TREC 2025 Proceedings

Submission Details

Organization

GenAIus

Track

RAG TREC Instrument for Multilingual Evaluation

Task

Report Generation Task

Date

2025-08-21

Run Description

Document collection

['English subset', 'Arabic subset', 'Chinese subset', 'Russian subset']

Machine translation of documents

['Yes we used the organizer-provided machine translations']

Write a short description of your retrieval process

(1) Generate questions using GPT-4o, (2) retrieve relevant documents for these questions using the provided search API

Write a short description of your generation process

(1) Generate nuggets from top retrieved documents, (2) cluster these nuggets, (3) use clusters to generate the final response

Which LLM(s) where used by your system?

GPT-4o

Open repository link

.

Assessing priority

2

Evaluation Files

• genaius-cluster.autoargue (autoargue)
• genaius-cluster.almost-human-judgments.tsv (almost-human-judgments.tsv)
• genaius-cluster.almost-human-scores.tsv (almost-human-scores.tsv)
• genaius-cluster.autoargue-scores.tsv (autoargue-scores.tsv)
• genaius-cluster.autoargue-judgments.jsonl (autoargue-judgments.jsonl)

Paper