TREC 2025 Proceedings

Submission Details

Organization

DUTH

Track

RAG TREC Instrument for Multilingual Evaluation

Task

Report Generation Task

Date

2025-08-21

Run Description

Document collection

['English subset']

Machine translation of documents

['None']

Write a short description of your retrieval process

-

Write a short description of your generation process

English-only report generation using our RRF retrieval run. For each topic we take top-20 documents, pick up to 6 high-overlap sentences (w.r.t. title+problem) and attach per-sentence citations. We add a short intro and enforce the per-topic character limit; “references” list the cited document IDs.

Which LLM(s) where used by your system?

None

Open repository link

-

Assessing priority

1 (highest)

Evaluation Files

• mlir-rrf-report.autoargue (autoargue)
• mlir-rrf-report.almost-human-judgments.tsv (almost-human-judgments.tsv)
• mlir-rrf-report.almost-human-scores.tsv (almost-human-scores.tsv)
• mlir-rrf-report.autoargue-scores.tsv (autoargue-scores.tsv)
• mlir-rrf-report.autoargue-judgments.jsonl (autoargue-judgments.jsonl)

Paper