TREC 2025 Proceedings

Submission Details

Organization

ncsu-las

Track

RAG TREC Instrument for Multilingual Evaluation

Task

Report Generation Task

Date

2025-08-14

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

track provided search service, plaidx

Write a short description of your generation process

changed the prompting, encouraged more iteration of reserach, and a split an agent into two different agents

Which LLM(s) where used by your system?

gpt-4.1 and gpt-4.1-mini

Open repository link

n/a

Assessing priority

2

Evaluation Files

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

Paper