Skip to content

Runs - RAG TREC Instrument for Multilingual Evaluation (RAGTIME) 2025

AMU1ENG

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: AMU1ENG
  • Participant: AMU
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 68120ad88832d1f40483aee801284347
  • Run description: The entire retrieval phase was performed on a locally hosted vector database. During the process of inserting texts into the database, they were divided into chunks of approximately 5 to 10 sentences, depending on their length. In edge cases, the maximum chunk size was set to 8k characters. These chunks were then vectorized using the base model BAAI/BGE-m3 with GPU acceleration. Retrieval was conducted using a standard similarity search based on the cosine similarity score.

AMU1ML

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: AMU1ML
  • Participant: AMU
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 015bdb5a7523380b80b12de017aff044
  • Run description: The entire retrieval phase was performed on a locally hosted vector database. During the process of inserting texts into the database, they were divided into chunks of approximately 5 to 10 sentences, depending on their length. In edge cases, the maximum chunk size was set to 8k characters. These chunks were then vectorized using the base model BAAI/BGE-m3 with GPU acceleration. Retrieval was conducted using a standard similarity search based on the cosine similarity score without reranking process

auto_swarm_mt

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: auto_swarm_mt
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: bac13ab176f8558e87f332098a9202b3
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

bm25-common-claims

Participants | Input | Appendix

  • Run ID: bm25-common-claims
  • Participant: EvalHLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 9eef99240cf869443854f40c798defe2
  • Run description: We use BM25 to retrieve the top 100 documents across languages, using the user story and the problem statement as the query

bm25-d-rank1

Participants | Input | trec_eval | Appendix

  • Run ID: bm25-d-rank1
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 6c8b1e95bfe78650b269204f630c5cb1
  • Run description: The queries (problem statement) were translated into each language of the collection subsets using google translate. Documents were indexed in their native language, separately, using BM25. The top 250 documents of each sub-collection was reranked with Rank-1. The 4 lists were merged by score.

bm25-d-rankk

Participants | Input | trec_eval | Appendix

  • Run ID: bm25-d-rankk
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 997229f2f8b635675a17f14a5f3f812b
  • Run description: The queries (problem statement) were translated into each language of the collection subsets using google translate. Documents were indexed in their native language, separately, using BM25. The top 250 documents of each sub-collection was reranked with Rank-1. The 4 lists were merged by score. The top 100 documents were reranked by Rank-k.

bm25-svc-claims

Participants | Input | Appendix

  • Run ID: bm25-svc-claims
  • Participant: EvalHLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: f74edee75208e3e13e67c4c19d429daf
  • Run description: We use the top 100 documents retrieved by BM25, using the user story and the problem statement as the query

bm25-t-rank1

Participants | Input | trec_eval | Appendix

  • Run ID: bm25-t-rank1
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: b4949f4af84fbf39f4959f03d735b045
  • Run description: The queries (title) were translated into each language of the collection subsets using google translate. Documents were indexed in their native language, separately, using BM25. The top 250 documents of each sub-collection was reranked with Rank-1. The 4 lists were merged by score.

bm25-t-rankk

Participants | Input | trec_eval | Appendix

  • Run ID: bm25-t-rankk
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: e5ef53733759725d2094bcfe77daac0c
  • Run description: The queries (title) were translated into each language of the collection subsets using google translate. Documents were indexed in their native language, separately, using BM25. The top 250 documents of each sub-collection was reranked with Rank-1. The 4 lists were merged by score. The top 100 documents were reranked by Rank-k.

bm25-td-rank1

Participants | Input | trec_eval | Appendix

  • Run ID: bm25-td-rank1
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 6921433c5e35d18b7263e6ca5f9c386f
  • Run description: The queries (title+problem statement) were translated into language each of the collection subsets using google translate. Documents were indexed in their native language, separately, using BM25. The top 250 documents of each sub-collection was reranked with Rank-1. The 4 lists were merged by score.

bm25-td-rankk

Participants | Input | trec_eval | Appendix

  • Run ID: bm25-td-rankk
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: e5ef53733759725d2094bcfe77daac0c
  • Run description: The queries (title+problem statement) were translated into language each of the collection subsets using google translate. Documents were indexed in their native language, separately, using BM25. The top 250 documents of each sub-collection was reranked with Rank-1. The 4 lists were merged by score. The top 100 documents were reranked by Rank-k.

bulleted_list

Participants | Input | Appendix

  • Run ID: bulleted_list
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 44602c8e71e9d91077d66e7d26a07c5d
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

bulleted_list_plus_rewrite

Participants | Input | Appendix

  • Run ID: bulleted_list_plus_rewrite
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: ca065d6f6b2dc60367c9f85d6de9dabd
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

chopped2000-crucible-nugget-references-most_common-SupportedAnswerExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-crucible-nugget-references-most_common-SupportedAnswerExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 8dac8adc310d3d3b1abb77642a688145
  • Run description: BM25

chopped2000-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerabilityExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerabilityExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 0b080da06975c0541ae36c36462ec976
  • Run description: top20 of ragtime_t+ps_lsr-listllama-ragtime.run

chopped2000-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: b380e41d0923a62dcacbd2845850cd4c
  • Run description: top20 of ragtime_t+ps_lsr-listllama-ragtime.run.

chopped2000-extractive-crucible-nugget-references-most_common-SupportedAnswerExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-extractive-crucible-nugget-references-most_common-SupportedAnswerExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: ae1654ed5491481900353da10923e299
  • Run description: BM25

chopped2000-strict-crucible-nugget-references-most_common-SupportedAnswerabilityExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-strict-crucible-nugget-references-most_common-SupportedAnswerabilityExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 7d3c3ac0aa79616ff982e02b0cbb3af3
  • Run description: BM25

chopped2000-strict-crucible-nugget-references-most_common-SupportedAnswerExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-strict-crucible-nugget-references-most_common-SupportedAnswerExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 4cd98b6fbf4783041839f392fef0c0e4
  • Run description: BM25

chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerabilityExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerabilityExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 5ad23a2623506f6dbf50623380f1c17c
  • Run description: top20 of ragtime_t+ps_lsr-listllama-ragtime.run.

chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_lsr-listllama-ragtime.run-SupportedAnswerExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 00d3cdb3e556af7bbf8fbea1a93a483a
  • Run description: top20 of ragtime_t+ps_lsr-listllama-ragtime.run.

chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_plaidx-listllama-ragtime.run-SupportedAnswerabilityExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_plaidx-listllama-ragtime.run-SupportedAnswerabilityExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 40f0fd724f4ef6f7ba79c00ec2fedeff
  • Run description: top20 of ragtime_t+ps_plaidx-listllama-ragtime.run.

chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_plaidx-listllama-ragtime.run-SupportedAnswerExtractor

Participants | Input | Appendix

  • Run ID: chopped2000-strict-crucible-ranking-most_common-ragtime_t+ps_plaidx-listllama-ragtime.run-SupportedAnswerExtractor
  • Participant: TREMA-UNH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: f151c1ee0a40b0aba1ba7b439da17d0b
  • Run description: top20 of ragtime_t+ps_plaidx-listllama-ragtime.run

cru-ablR-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ablR-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: ea4b5be423955fbccaedff41d159caef
  • Run description: See below

cru-ablR-conf-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ablR-conf-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 797146666bab8aa1ef0cb6411bf9560f
  • Run description: See below

cru-ablR-LSR-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ablR-LSR-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: d21f545a4408fbea28f30398616435f5
  • Run description: See below

cru-ablR-PlaidX-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ablR-PlaidX-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: cbd4851cb80030df54db55f9ca2040a2
  • Run description: See below

cru-ansR-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ansR-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 9e8e0066b82fc7e912fb3171f458e211
  • Run description: See below

cru-ansR-bareconf-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ansR-bareconf-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: cc0da5220f5d5938468f6b7ab9e5045b
  • Run description: See below

cru-ansR-conf-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ansR-conf-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 8a5ca874702cede490c3ca22880730eb
  • Run description: See below

cru-ansR-LSR-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ansR-LSR-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: d11467f60a276a2fdf746b5ad89fe15d
  • Run description: See below

cru-ansR-mostcommon-

Participants | Proceedings | Input | almost-human-scores.tsv | almost-human-judgments.tsv | autoargue | Appendix

  • Run ID: cru-ansR-mostcommon-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 3d3a931c812f42496c46aa3fd1728c44
  • Run description: See below

cru-ansR-PlaidX-

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: cru-ansR-PlaidX-
  • Participant: HLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: ea811c913cd407056f59a1b23518b738
  • Run description: See below

dfki-milp-base

Participants | Input | autoargue | Appendix

  • Run ID: dfki-milp-base
  • Participant: DFKI
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: ad5d45aae7cc1e567b3f79b10c8bf89a
  • Run description: Plaid search API

dfki-milp-baseline

Participants | Input | Appendix

  • Run ID: dfki-milp-baseline
  • Participant: dfki
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 3e17a8244e574f18cd5a01af6c7f9565
  • Run description: The retrieval works via plaid search API

dry_all3_eng_1000

Participants | Proceedings | Input | Appendix

  • Run ID: dry_all3_eng_1000
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 6ebf49bc0e041f5861fedd915a3c24e4
  • Run description: All three fields for retrieval

dry_gq_eng_1000

Participants | Proceedings | Input | Appendix

  • Run ID: dry_gq_eng_1000
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: c1df50a77fbbf90a7cb893deea446030
  • Run description: Retrieval of questions

duth-mlir-mlm6

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: duth-mlir-mlm6
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: e64118147d03f65600ed7e9e96e9d2c5
  • Run description: English-only dense retrieval using a MiniLM 6-layer encoder on the RAGTIME1 English subset. We index document embeddings and rank by vector similarity; outputs are deduplicated and cut to top-1000.

duth-mlir-mlm6loc

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: duth-mlir-mlm6loc
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 0d0b2b442bbb2681c509c23cab6c0a24
  • Run description: English-only dense retrieval with a MiniLM 6-layer local variant (same model family, local indexing/hyperparameters) on the RAGTIME1 English subset. We rank by vector similarity and return the top-1000 per topic.

duth_mlir_eng_rrf

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: duth_mlir_eng_rrf
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 2d099cbf23eecd4548f9c531566d8a79
  • Run description: English-only retrieval on the RAGTIME1 English subset. We combine BM25 with multiple multilingual dense retrievers (x-encoder, MiniLM-6/12, ELECTRA, TinyBERT) using standard Reciprocal Rank Fusion (RRF). Per-topic lists are deduplicated and cut to top-1000.

duth_mlir_xenc

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: duth_mlir_xenc
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 54e7a12652a2dc0b9d056d582eee7829
  • Run description: English-only retrieval with a cross-lingual transformer encoder (“x-encoder”) for dense retrieval over the RAGTIME1 English subset. Results are scored by dot-product similarity, deduplicated, and truncated to top-1000 per topic.

electra

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: electra
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-22
  • Task: trec2025-ragtime-repgen
  • MD5: 13bf500fd28b73ac7550a96c0e2c4600
  • Run description: ELECTRA encoder dense retrieval; vector similarity; dedup + top-1000.

embed.mt_q3d3

Participants | Input | Appendix

  • Run ID: embed.mt_q3d3
  • Participant: hltcoe-cascade
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 5b8df0a393c1ca92c382238453d7f2f8
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

eng-IITJ-faiss_bge_v4-run4

Participants | Input | Appendix

  • Run ID: eng-IITJ-faiss_bge_v4-run4
  • Participant: IITJ
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 00ee682f534411c326d9a904cb26ee11
  • Run description: We used a dense retrieval approach based on the BAAI/bge-base-en-v1.5 sentence transformer. The corpus was encoded into dense embeddings and indexed using FAISS with inner product similarity. Queries were similarly encoded and normalized before performing top-k similarity search over the FAISS index. The results were formatted in TREC run file format for evaluation.

eng-IITJ-faiss_bge_v5-run5

Participants | Input | Appendix

  • Run ID: eng-IITJ-faiss_bge_v5-run5
  • Participant: IITJ
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 7961c142594976df603abad5018cfdb3
  • Run description: We used a dense retrieval approach based on the BAAI/bge-base-en-v1.5 sentence transformer. The corpus was encoded into dense embeddings and indexed using FAISS with inner product similarity. Queries were similarly encoded and normalized before performing top-k similarity search over the FAISS index. The results were formatted in TREC run file format for evaluation.

eng-IITJ-faiss_bge_v6-run6

Participants | Input | Appendix

  • Run ID: eng-IITJ-faiss_bge_v6-run6
  • Participant: IITJ
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: e5be718a8962f4533ff8bfa12587d7d1
  • Run description: We used a dense retrieval approach based on the BAAI/bge-base-en-v1.5 sentence transformer. The corpus was encoded into dense embeddings and indexed using FAISS with inner product similarity. Queries were similarly encoded and normalized before performing top-k similarity search over the FAISS index. The results were formatted in TREC run file format for evaluation.

eng-IITJ-faiss_bge_v7-run7

Participants | Input | Appendix

  • Run ID: eng-IITJ-faiss_bge_v7-run7
  • Participant: IITJ
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 9fcefdc31d325c18af0d784346ada322
  • Run description: We used a dense retrieval approach based on the BAAI/bge-base-en-v1.5 sentence transformer. The corpus was encoded into dense embeddings and indexed using FAISS with inner product similarity. Queries were similarly encoded and normalized before performing top-k similarity search over the FAISS index. The results were formatted in TREC run file format for evaluation.

eng-IITJ-faiss_bge_v8-run8

Participants | Input | Appendix

  • Run ID: eng-IITJ-faiss_bge_v8-run8
  • Participant: IITJ
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: fcec05bcd2615832adea0a1d11de70af
  • Run description: We used a dense retrieval approach based on the BAAI/bge-base-en-v1.5 sentence transformer. The corpus was encoded into dense embeddings and indexed using FAISS with inner product similarity. Queries were similarly encoded and normalized before performing top-k similarity search over the FAISS index. The results were formatted in TREC run file format for evaluation.

eng-IITJ-faiss_bge_v9-run9

Participants | Input | Appendix

  • Run ID: eng-IITJ-faiss_bge_v9-run9
  • Participant: IITJ
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: acec0f2c84e09bebf8b3000e02c1ba44
  • Run description: We used a dense retrieval approach based on the BAAI/bge-base-en-v1.5 sentence transformer. The corpus was encoded into dense embeddings and indexed using FAISS with inner product similarity. Queries were similarly encoded and normalized before performing top-k similarity search over the FAISS index. The results were formatted in TREC run file format for evaluation.

eng_fused

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: eng_fused
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: b29915f6d97cf1c9b19b620f2731e53d
  • Run description: English-only report generation using our fused retrieval run (linear combination). Same extractive pipeline with per-sentence citations and character-limit.

eng_mlm6

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: eng_mlm6
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 9de1abf1bd22fc901312e1a022ccd7c9
  • Run description: English-only report generation using our MiniLM-6 retrieval run. Top-20 docs, up to 6 high-overlap sentences with per-sentence citations, short intro, character-limit enforced; references list cited document IDs

eng_mlm6loc

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: eng_mlm6loc
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 60ade1441fdfe7e7ad23322b90099fc1
  • Run description: English-only report generation using our MiniLM-6 local variant retrieval run. Same pipeline as above.

extractive-rag

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: extractive-rag
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 56fccf12df6611dd5ab3f909e205bc34
  • Run description: This run uses the track-provided PLAID-X search service.

genaius-cluster

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: genaius-cluster
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 518c8183d7eeb974a8fed3876cbfb816
  • Run description: (1) Generate questions using GPT-4o, (2) retrieve relevant documents for these questions using the provided search API

genaius-gpt-4o

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: genaius-gpt-4o
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 023baccb628cbf2f92f25f41a71e680f
  • Run description: (1) Generate questions using GPT-4o, (2) retrieve relevant documents for these questions using the provided search API, (3) merge the retrieved documents

genaius-gpt-oss-120b

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: genaius-gpt-oss-120b
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: cd976d1583fb1811342fc63e8439a0b3
  • Run description: (1) Generate questions using gpt-oss-120b, (2) retrieve relevant documents for these questions using the provided search API, (3) merge the retrieved documents

genaius-gpt-oss-20b

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: genaius-gpt-oss-20b
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: e82f3fe7f6c790fd3b8759b270435732
  • Run description: (1) Generate questions using gpt-oss-20b, (2) retrieve relevant documents for these questions using the provided search API, (3) merge the retrieved documents

genaius-llama3-3-70B

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: genaius-llama3-3-70B
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: e508e0d33089821f7cfce5ec4bf4e422
  • Run description: (1) Generate questions using Llama3.3-70B, (2) retrieve relevant documents for these questions using the provided search API, (3) merge the retrieved documents

genaius-question

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: genaius-question
  • Participant: GenAIus
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 7c126d304e6be2105442835ea14db58e
  • Run description: (1) Generate questions using GPT-4o, (2) retrieve relevant documents for these questions using the provided search API

goldenish-reports-20-most-common.jsonl

Participants | Input | Appendix

  • Run ID: goldenish-reports-20-most-common.jsonl
  • Participant: EvalHLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 100c80eca9e8df836898621d66c26da7
  • Run description: We retrieve the top 100 documents for each report request using BM25

gptr_e2_q3d3_mt

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: gptr_e2_q3d3_mt
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: da592ba81198a44803c30b2e0beed83e
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

gptr_ka_q3d3_mt

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: gptr_ka_q3d3_mt
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 474c78b382b86883f4b30c1cebfe0131
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

gptr_ka_q3d3_natv

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: gptr_ka_q3d3_natv
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 63c99b2a8d5c93995b83659a52eb82be
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

gptr_nt_q3d3_mt

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: gptr_nt_q3d3_mt
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 49d0fcade5d1916eceb6eb4cf8b7e2a7
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

gptr_nt_q4d4_mt

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: gptr_nt_q4d4_mt
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 15c8a190d9d68d6f4498937e93c03b76
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

high-to-low-interview-1-rrf

Participants | Input | Appendix

  • Run ID: high-to-low-interview-1-rrf
  • Participant: hltcoe-cascade
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 546083518915b840b2c983dead7b6324
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

high-to-low-interview-3-rrf

Participants | Input | Appendix

  • Run ID: high-to-low-interview-3-rrf
  • Participant: hltcoe-cascade
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 93326fca65766d9cd6794c3b55405c84
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

hltcoe-rerank-crux

Participants | Input | Appendix

  • Run ID: hltcoe-rerank-crux
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: db024172fc371cd142e922499d8ef7ce
  • Run description: Reranking RRF for question coverage with CRUX using Llama-3.3-70B-Instruct

hltcoe-rerank-knn

Participants | Input | Appendix

  • Run ID: hltcoe-rerank-knn
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 56bc6925fcbf61316ac8ac61bc13e40e
  • Run description: Corpus graph KNN reranking on top of PLAID-X

hltcoe-rerank-listllama-fusion3

Participants | Input | Appendix

  • Run ID: hltcoe-rerank-listllama-fusion3
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 98aa6efe9c96108a27a889412912c998
  • Run description: Listwise reranking using Llama3.3-70B-Instruct

hltcoe-rerank-searcher

Participants | Input | Appendix

  • Run ID: hltcoe-rerank-searcher
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 0dcd8154ca3be297ca51dd53c0ec99f5
  • Run description: Searcher II reranking RRF Qwen3, LSR, PLAID-X

hltcoe-retrieve-expand_splade_max75

Participants | Input | Appendix

  • Run ID: hltcoe-retrieve-expand_splade_max75
  • Participant: hltcoe-retrieve
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 1f94af55d30ad7d17ac67652c0a368f3
  • Run description: Query decomp with LSR

hltcoe-retrieve-lsr

Participants | Input | Appendix

  • Run ID: hltcoe-retrieve-lsr
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 084c71bc4f63829b208cfe340cb96448
  • Run description: MultiLSR

hltcoe-retrieve-maxsimmmr

Participants | Input | Appendix

  • Run ID: hltcoe-retrieve-maxsimmmr
  • Participant: hltcoe-retrieve
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 1ab6c60ab2cde59ab07beb1acb08d691
  • Run description: MaxSim MMR with PLAID-X

hltcoe-retrieve-plaidx-decomp

Participants | Input | Appendix

  • Run ID: hltcoe-retrieve-plaidx-decomp
  • Participant: hltcoe-retrieve
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: dcc32fa86a4bbbe03ecddfc3abb25df6
  • Run description: Query decomp with PLAID-X

hltcoe-retrieve-plaidx_lsr_qwen

Participants | Input | Appendix

  • Run ID: hltcoe-retrieve-plaidx_lsr_qwen
  • Participant: hltcoe-retrieve
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 4c87eb577d57f2396a562f31a86c99d0
  • Run description: Rank fusion of Qwen3, LSR, PLAID-X

hltcoe-retrieve-qwen

Participants | Input | Appendix

  • Run ID: hltcoe-retrieve-qwen
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: b840818e2a3103452f0db99ed24ffe7b
  • Run description: Dense retrieval using Qwen3-8B embeddings

hltime-fsqwen

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-fsqwen
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: ee5e77b81a002563737cbfb6a4973847
  • Run description: Qwen-Embedding-8B. Translations were not used.

hltime-fsrrf

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-fsrrf
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 5a647deda1f3bf3e28cc7e19886d0b85
  • Run description: RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used.

hltime-fsrrfprf

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-fsrrfprf
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 3b58b9f8aff357b1dd494283e75946ad
  • Run description: RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR using PRF with each. Translations were not used.

hltime-gpt5.searcher

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-gpt5.searcher
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 141c5f04fa64895b0b6ec26b3994e420
  • Run description: Searcher II pointwise reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR.

hltime-lg.crux

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.crux
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 486d53d4c9861c2762daf51b526c6b1e
  • Run description: Llama-3.3-70B-Instruct pointwise subquestion reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-lg.fsrrf

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.fsrrf
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 4443965248feb37cba3b6c677a010881
  • Run description: RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-lg.fsrrfprf

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.fsrrfprf
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 0d2cf9cb9f0131a4270b81b505ae7883
  • Run description: RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR using PRF with each. Translations were not used.

hltime-lg.jina

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.jina
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 2838f0b6bad833de8048f85764e887e0
  • Run description: jinaai/jina-reranker-m0 (2.4B) reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-lg.jina.qwen

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.jina.qwen
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 7125f656fa34eef05b24891ba9622a72
  • Run description: RRF of jinaai/jina-reranker-m0 (2.4B) and Qwen/Qwen3-Reranker-8B reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-lg.listllama

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.listllama
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 4d827ba35dab1252143b11dc5518117f
  • Run description: Llama-3.3-70B-Instruct listwise reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-lg.qwen

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.qwen
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: c92c0097ff37e0882fd9b61f3cdb72fd
  • Run description: Qwen/Qwen3-Reranker-8B reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-lg.searcher

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: hltime-lg.searcher
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 7f6a80a7a51e7963cc6f8fcd3f27362b
  • Run description: Searcher II pointwise reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were not used for first-stage retrieval.

hltime-listllama

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-listllama
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 8b06a3deac7934879444454780feea88
  • Run description: Llama-3.3-70B-Instruct listwise reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were used for reranking but not used for first-stage retrieval.

hltime-lsr

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-lsr
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 8fe7f7d514da1d68c729c12c44742b34
  • Run description: Multilingual LSR using XLM-Roberta-Large. Translations were not used.

hltime-plaidx

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-plaidx
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 51c94f948588e16e4f9210ec2b2dc5e6
  • Run description: PLAID-X. Translations were not used.

hltime-qwen

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-qwen
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 8d885291dccb5e5d89e7d2cbd8f3edf5
  • Run description: Qwen/Qwen3-Reranker-8B reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were used for reranking but not used for first-stage retrieval.

hltime-qwen-jina

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-qwen-jina
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 59ac0fa476019550c6d4764e81cbb1ae
  • Run description: Fusion of jina-reranker-m0 and Qwen-Embedding-8B reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were used for reranking but not used for first-stage retrieval.

hltime-rankk

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-rankk
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: efd1fe0d7b92a7d3bb56a10bb2290331
  • Run description: Rank-K reranking the top 100 from Rank-1, which reranked the top 1000 from RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were used for reranking but not used for first-stage retrieval.

hltime-searcher

Participants | Input | trec_eval | Appendix

  • Run ID: hltime-searcher
  • Participant: hltcoe-rerank
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: cfc5c0e137a83cf8890a68b347bd5ee7
  • Run description: Searcher II pointwise reranking RRF of PLAID-X, Qwen-Embedding-8B, and multilingual LSR. Translations were used for reranking but not used for first-stage retrieval.

IDACCS_extract_4.1

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: IDACCS_extract_4.1
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 0f74307d2f6a69b5557fccc290a91f6f
  • Run description: \item The organizers served to retrieve the top 30 documents using the background and problem statement as a query. \item We reranked the documents to get the top 10 using \texttt{mxbai-rerank-large-v1} on 10 sentence chunks with an overlap of 5, using a query generated by GPT-4o based on the title, background, and problem statement.

IDACCS_hybrid_4.1

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: IDACCS_hybrid_4.1
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-19
  • Task: trec2025-ragtime-repgen
  • MD5: a22cfa5849cf8ea237d4fd9675353555
  • Run description: The following steps were done.
  • The organizers serve to retrieve the top 30 documents using the background and problem statement as a query.
  • We reranked the document to get the top 10 using mxbai-rerank-large-v1 on 10 sentence chunks with an overlap of 5 using a query generated by gpt-4o base on the title, background, and problem statement.

IDACCS_hybridtb_4.1

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: IDACCS_hybridtb_4.1
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 14e503c8dcd68d8a43d67bb35bbb20ac
  • Run description: \item The organizers served to retrieve the top 30 documents using the background and problem statement as a query. \item We reranked the documents to get the top 10 using \texttt{mxbai-rerank-large-v1} on 10 sentence chunks with an overlap of 5, using a query generated by title, and background

IDACCS_nugget_4.1

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: IDACCS_nugget_4.1
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-19
  • Task: trec2025-ragtime-repgen
  • MD5: 05b271048dc1c310c52e2cb8d7b87f65
  • Run description: The following steps were done.
  • The organizers serve to retrieve the top 30 documents using the background and problem statement as a query.
  • We reranked the document to get the top 10 using mxbai-rerank-large-v1 on 10 sentence chunks with an overlap of 5 using a query generated by gpt-4o base on the title, background, and problem statement.

IDACCS_nugget_tb4.1

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: IDACCS_nugget_tb4.1
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: b671ddded7afab9d643eb0af30886122
  • Run description: \item The organizers served to retrieve the top 30 documents using the background and problem statement as a query. \item We reranked the documents to get the top 10 using \texttt{mxbai-rerank-large-v1} on 10 sentence chunks with an overlap of 5, using a query generated by GPT-4o based on the title, background, and problem statement.

keep_all.mt_q3d3

Participants | Input | Appendix

  • Run ID: keep_all.mt_q3d3
  • Participant: hltcoe-cascade
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 19b3d73c057313780f1bf0b50c3bc33e
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

keep_all.src_q3d3

Participants | Input | Appendix

  • Run ID: keep_all.src_q3d3
  • Participant: hltcoe-cascade
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 6f23648ba5cd6e69071f667fb38d4a4c
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

las_ag_round_robin

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: las_ag_round_robin
  • Participant: ncsu-las
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-repgen
  • MD5: d3b2187a0544afc1eddc13d7d34f23e1
  • Run description: track provided plaidx one

las_ag_sel_28

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: las_ag_sel_28
  • Participant: ncsu-las
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-repgen
  • MD5: 41d0922c2f32227ef9b4ec5e10198da8
  • Run description: track provided search service

las_ag_sel_29

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: las_ag_sel_29
  • Participant: ncsu-las
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-repgen
  • MD5: 7b2430be8403361358a23d39bd273882
  • Run description: track provided search service

las_ag_sel_all_4.1

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: las_ag_sel_all_4.1
  • Participant: ncsu-las
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-repgen
  • MD5: 5a5ee9567fbaec439b96746ac7ad76e2
  • Run description: track provided service

las_ag_sel_new_prompt

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: las_ag_sel_new_prompt
  • Participant: ncsu-las
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-repgen
  • MD5: 167c57f70232c88ef6d4bcfd5e4c714b
  • Run description: track provided search service, plaidx

lg-w-ret_lq-nt-s_cite-4q.3l.5r

Participants | Input | Appendix

  • Run ID: lg-w-ret_lq-nt-s_cite-4q.3l.5r
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: e16813861191f7d79965b8a9a76cab93
  • Run description: Rank fusion of Qwen3-8B Embedding model and Multilingual LSR using XLM-RoBERTa-Large

lg-w-ret_plq-nt-s_cite-4q.3l.5r

Participants | Input | Appendix

  • Run ID: lg-w-ret_plq-nt-s_cite-4q.3l.5r
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: b0e7950e2bd0584d3f8dca695e7ac3a4
  • Run description: Rank fusion of PLAID-X, Qwen3-8B Embedding model, and Multilingual LSR using XLM-RoBERTa-Large

lg_e2_3q5r2l_mt_qw3

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: lg_e2_3q5r2l_mt_qw3
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 858997418aa0d962b963ff3a49eb138d
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

lg_e2_3q5r3l

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: lg_e2_3q5r3l
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 7fd7a3e230377a3b6af99aa5bc13f4b6
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

lg_nt_4q12r3l_mt_c

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: lg_nt_4q12r3l_mt_c
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 7d21a61f6e4124fc99449b8b2b3584f6
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

lg_nt_4q12r3l_natv_c

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: lg_nt_4q12r3l_natv_c
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 41ad3ec9122ad1052bce81d3bc13a124
  • Run description: Reciprocal ranked fusion over multilingual retrieval using plaid-x, multilingual LSR, and Qwen3 single dense vectors with ANN search.

lsr_qwen.concise_prompts.2loop.7docs

Participants | Input | Appendix

  • Run ID: lsr_qwen.concise_prompts.2loop.7docs
  • Participant: hltcoe-multiagt
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 6d024d9ef62c1697cbea2d41e4c75c22
  • Run description: Rank fusion of Qwen3-8B Embedding model and Multilingual LSR using XLM-RoBERTa-Large

milp-query-expanded

Participants | Input | autoargue | Appendix

  • Run ID: milp-query-expanded
  • Participant: DFKI
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 2f99d37b2694524def84cdf9cc5291d6
  • Run description: PLAID search API

mlir-elec

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: mlir-elec
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: f1a94ae653f4bc7948c21c300dda54bf
  • Run description: English-only dense retrieval with an ELECTRA encoder on the RAGTIME1 English subset. We rank by vector similarity and keep the top-1000 per topic after deduplication.

mlir-fused

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: mlir-fused
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 5610ffd89479843c73e6bace5bbbe292
  • Run description: English-only linear fusion of BM25 with selected dense runs (score-normalized then combined) on the RAGTIME1 English subset. The merged list is deduplicated and truncated to top-1000 per topic.

mlir-hltcoe-METS-dt-title+desc.spacy.bm25.rm3

Participants | Input | Appendix

  • Run ID: mlir-hltcoe-METS-dt-title+desc.spacy.bm25.rm3
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: e70459863cfbf88e2d8f8ba59cb184ed
  • Run description: BM25 with RM3 using the title+problem statement as the query

mlir-mlm12

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: mlir-mlm12
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 0ece8c74e0885ca9e5c5cf90f9d9d84d
  • Run description: English-only dense retrieval using a MiniLM 12-layer encoder on the RAGTIME1 English subset. Documents are ranked by embedding similarity; lists are deduplicated and cut to top-1000.

mlir-pybm25

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: mlir-pybm25
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 54e7a12652a2dc0b9d056d582eee7829
  • Run description: English-only BM25 baseline on the RAGTIME1 English subset (standard parameters). We output a single ranked list per topic, deduplicated and truncated to top-1000.

mlir-rrf-report

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: mlir-rrf-report
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 54efefb30a3c6a917bd2d32f0c60faf2
  • Run description: -

mlir-tb

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: mlir-tb
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 3acec6ef1492459fbe287384c10f9e78
  • Run description: English-only dense retrieval with TinyBERT on the RAGTIME1 English subset. We compute embedding similarity for ranking, deduplicate, and return top-1000 per topic.

mlir-tblocal

Participants | Proceedings | Input | trec_eval | Appendix

  • Run ID: mlir-tblocal
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 12395a12590c0151fbbeb8f7074bfbc0
  • Run description: English-only dense retrieval with a TinyBERT local variant on the RAGTIME1 English subset. We produce a single ranked list per topic by vector similarity, deduplicate, and keep top-1000.

mlm12

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: mlm12
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-22
  • Task: trec2025-ragtime-repgen
  • MD5: 84dad6f73064c955eaa5e6cb4a50dad4
  • Run description: MiniLM-12 dense retrieval; vector similarity; dedup + top-1000.

mt-bm25-td

Participants | Input | trec_eval | Appendix

  • Run ID: mt-bm25-td
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: cf30002ea6579eda822825a55444c555
  • Run description: The English translations of the none English documents, along with the English documents, were indexed with BM25. The title and problem statement were used as the query. Ranking included rm3.

mt-bm25-title

Participants | Input | trec_eval | Appendix

  • Run ID: mt-bm25-title
  • Participant: coordinators
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-mlir
  • MD5: 248caff75a3680e1a23310df05a74e09
  • Run description: The English translations of the none English documents, along with the English documents, were indexed with BM25. The title was used as the query. Ranking included rm3.

plaidx-merged-questions

Participants | Input | Appendix

  • Run ID: plaidx-merged-questions
  • Participant: EvalHLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 3f780962140777f7324797debd5323cd
  • Run description: We use PLAID-X to retrieve the top 100 documents, using the user story and the problem statement as the query

plaidx-svc-questions

Participants | Input | Appendix

  • Run ID: plaidx-svc-questions
  • Participant: EvalHLTCOE
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: a81ba97cf3a502b0c0fb0d69451f7049
  • Run description: We use PLAID-X to retrieve the top 100 documents across languages, using the user story and the problem statement as the query

plaidx.title-prostat_merged

Participants | Input | Appendix

  • Run ID: plaidx.title-prostat_merged
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: ddd19bc44797d4af313444176b1c7eb5
  • Run description: MTD PLAID-X retrieval with score fusion on separate indexes

pybm25

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: pybm25
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-22
  • Task: trec2025-ragtime-repgen
  • MD5: 3bb4628c92d27a2e3d99ecad071fbfc8
  • Run description: BM25 baseline; single ranked list; dedup + top-1000.

query_expanded+milp-KPE

Participants | Input | Appendix

  • Run ID: query_expanded+milp-KPE
  • Participant: dfki
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 2f99d37b2694524def84cdf9cc5291d6
  • Run description: plaid-search

ragtime_run_100_topics_25_docs_per_query_per_languge_1_rus_arb_zho_eng_20250708_14_33_occams_budget_1000_hybrid_10_reranked

Participants | Proceedings | Input | Appendix

  • Run ID: ragtime_run_100_topics_25_docs_per_query_per_languge_1_rus_arb_zho_eng_20250708_14_33_occams_budget_1000_hybrid_10_reranked
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 6e44f01da6949b2d90b6c9faf43d3082
  • Run description: Stratified retrival across the four languages

ragtime_run_100_topics_25_docs_per_query_per_languge_1_rus_arb_zho_eng_20250708_14_33_occams_budget_1000_nuggets_10_reranked

Participants | Proceedings | Input | Appendix

  • Run ID: ragtime_run_100_topics_25_docs_per_query_per_languge_1_rus_arb_zho_eng_20250708_14_33_occams_budget_1000_nuggets_10_reranked
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 9c31051180bc13f86593436e01658fb1
  • Run description: Stratified sample across four languages

ragtime_run_75_topics_25_docs_per_query_per_languge_1_20250707_00_08_occams_budget_4000_hybrid_raranked_concat_by_lang

Participants | Proceedings | Input | Appendix

  • Run ID: ragtime_run_75_topics_25_docs_per_query_per_languge_1_20250707_00_08_occams_budget_4000_hybrid_raranked_concat_by_lang
  • Participant: IDACCS
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: a4d66f862cca964e88c124f5d8a889b1
  • Run description: Stratified retrieval across three languages using organizers server.

rankk100_lsr_qwen

Participants | Input | Appendix

  • Run ID: rankk100_lsr_qwen
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 99ba28dc323d76ba104085a78e0f1e02
  • Run description: Rank-K reranking top 100 from RRF of Qwen3 and MultiLSR

rankk100_plaidx

Participants | Input | Appendix

  • Run ID: rankk100_plaidx
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 7b7a8b8173ce881156345bbf0541fcf0
  • Run description: RankK reranking top 100 from MTD PLAID-X

rankk100_plaidx_lsr_qwen

Participants | Input | Appendix

  • Run ID: rankk100_plaidx_lsr_qwen
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 60fa2d28800520b1a36013d6b26ea49c
  • Run description: RankK reranking top 100 from RRF of Qwen3, Multi-LSR, and MTD PLAID-X

rankk200.rank1.native_title_desc_spacy_rm3_merged

Participants | Input | Appendix

  • Run ID: rankk200.rank1.native_title_desc_spacy_rm3_merged
  • Participant: hltcoe
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: 3b5e3ad270a7999c1b6586096bdc04eb
  • Run description: Rank1 then Rank-K on BM25 with title+problem statement

tb

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: tb
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-22
  • Task: trec2025-ragtime-repgen
  • MD5: 42d4c05663d34768ddf465e98a3fcad4
  • Run description: TinyBERT dense retrieval; vector similarity; dedup + top-1000.

tblocal

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: tblocal
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-22
  • Task: trec2025-ragtime-repgen
  • MD5: 3da3f5a09f033ac1a63b6a306b16bcab
  • Run description: TinyBERT (local variant); vector similarity; dedup + top-1000.

v1_qwen

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: v1_qwen
  • Participant: CSU
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 13eee22d4a391185ebc9311f746c9899
  • Run description: The system uses pre-built FAISS indexes containing document embeddings. When a query arrives, it is encoded using a multilingual BGE model with language-specific prefixes. The encoded query vector searches the indexes to retrieve top-k similar documents from multiple language splits. Results are filtered by a minimum similarity score threshold to ensure relevance.

v2_split_qwen

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: v2_split_qwen
  • Participant: CSU
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: 708d319bf2db9cf9d391538d8be55b37
  • Run description: The system uses pre-built FAISS indexes containing document embeddings. When a query arrives, it is encoded using a multilingual BGE model with language-specific prefixes. The encoded query vector searches the indexes to retrieve top-k similar documents from multiple language splits. Results are filtered by a minimum similarity score threshold to ensure relevance.

v3_surround_glm4

Participants | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: v3_surround_glm4
  • Participant: CSU
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-20
  • Task: trec2025-ragtime-repgen
  • MD5: a532f774f8b008964a57a3cbf9ae9131
  • Run description: The system uses pre-built FAISS indexes containing document embeddings. When a query arrives, it is encoded using a multilingual BGE model with language-specific prefixes. The encoded query vector searches the indexes to retrieve top-k similar documents from multiple language splits. Results are filtered by a minimum similarity score threshold to ensure relevance.

WueRAG_2025_07_08_20_05_00

Participants | Proceedings | Input | Appendix

  • Run ID: WueRAG_2025_07_08_20_05_00
  • Participant: WueRAG
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-14
  • Task: trec2025-ragtime-dryrun
  • MD5: a83e22a5657f679e2d72eb850722a33e
  • Run description: For our first submission we used a basic embedding distance similarity for our retrieval. We plan to iterate on this four our final submission. We used sentence-transformers/all-MiniLM-L6-v2 as embedding model.

WueRAG_2025_08_22

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: WueRAG_2025_08_22
  • Participant: WueRAG
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-22
  • Task: trec2025-ragtime-repgen
  • MD5: aa8b04eb86c3a89ccc4e8b3c8db73dfb
  • Run description: The retrieval process starts by reformulating the user query for better search relevance. Depending on configuration, the system either queries a remote RAGTIME API or searches local indices (for the last submission the RAGTIME API was used with subset set to "eng"). Retrieved documents are converted into a temporary index and processed via an embedding-based and BM25 retriever, in an hybrid setting.

xenc-report

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: xenc-report
  • Participant: DUTH
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-21
  • Task: trec2025-ragtime-repgen
  • MD5: 90bfeeda8f5565a9f49ede2c048dfcfe
  • Run description: -

zetaalpha

Participants | Proceedings | Input | almost-human-scores.tsv | autoargue | autoargue-scores.tsv | autoargue-judgments.jsonl | almost-human-judgments.tsv | Appendix

  • Run ID: zetaalpha
  • Participant: UvA
  • Track: RAG TREC Instrument for Multilingual Evaluation (RAGTIME)
  • Year: 2025
  • Submission: 2025-08-26
  • Task: trec2025-ragtime-repgen
  • MD5: af7016132dec87ab78758b152e91a82a
  • Run description: sets of document get read by llm as a judge which extracts factsoids.