Saved in:
Bibliographic Details
Main Authors: Li, Xiaozhe, Lyu, Tianyi, Yang, Siyi, Yang, Yizhao, Gong, Yuxi, Huang, Jinxuan, Zhang, Ligao, Huang, Zhuoyi, Liu, Qingwen
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.21329
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910063030960128
author Li, Xiaozhe
Lyu, Tianyi
Yang, Siyi
Yang, Yizhao
Gong, Yuxi
Huang, Jinxuan
Zhang, Ligao
Huang, Zhuoyi
Liu, Qingwen
author_facet Li, Xiaozhe
Lyu, Tianyi
Yang, Siyi
Yang, Yizhao
Gong, Yuxi
Huang, Jinxuan
Zhang, Ligao
Huang, Zhuoyi
Liu, Qingwen
contents Understanding human intent is a high-level cognitive challenge for Large Language Models (LLMs), requiring sophisticated reasoning over noisy, conflicting, and non-linear discourse. While LLMs excel at following individual instructions, their ability to distill Collective Intent - the process of extracting consensus, resolving contradictions, and inferring latent trends from multi-source public discussions - remains largely unexplored. To bridge this gap, we introduce COIN-BENCH, a dynamic, real-world, live-updating benchmark specifically designed to evaluate LLMs on collective intent understanding within the consumer domain. Unlike traditional benchmarks that focus on transactional outcomes, COIN-BENCH operationalizes intent as a hierarchical cognitive structure, ranging from explicit scenarios to deep causal reasoning. We implement a robust evaluation pipeline that combines a rule-based method with an LLM-as-the-Judge approach. This framework incorporates COIN-TREE for hierarchical cognitive structuring and retrieval-augmented verification (COIN-RAG) to ensure expert-level precision in analyzing raw, collective human discussions. An extensive evaluation of 20 state-of-the-art LLMs across four dimensions - depth, breadth, informativeness, and correctness - reveals that while current models can handle surface-level aggregation, they still struggle with the analytical depth required for complex intent synthesis. COIN-BENCH establishes a new standard for advancing LLMs from passive instruction followers to expert-level analytical agents capable of deciphering the collective voice of the real world. See our project page on COIN-BENCH.
format Preprint
id arxiv_https___arxiv_org_abs_2603_21329
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle COINBench: Moving Beyond Individual Perspectives to Collective Intent Understanding
Li, Xiaozhe
Lyu, Tianyi
Yang, Siyi
Yang, Yizhao
Gong, Yuxi
Huang, Jinxuan
Zhang, Ligao
Huang, Zhuoyi
Liu, Qingwen
Information Retrieval
Artificial Intelligence
Understanding human intent is a high-level cognitive challenge for Large Language Models (LLMs), requiring sophisticated reasoning over noisy, conflicting, and non-linear discourse. While LLMs excel at following individual instructions, their ability to distill Collective Intent - the process of extracting consensus, resolving contradictions, and inferring latent trends from multi-source public discussions - remains largely unexplored. To bridge this gap, we introduce COIN-BENCH, a dynamic, real-world, live-updating benchmark specifically designed to evaluate LLMs on collective intent understanding within the consumer domain. Unlike traditional benchmarks that focus on transactional outcomes, COIN-BENCH operationalizes intent as a hierarchical cognitive structure, ranging from explicit scenarios to deep causal reasoning. We implement a robust evaluation pipeline that combines a rule-based method with an LLM-as-the-Judge approach. This framework incorporates COIN-TREE for hierarchical cognitive structuring and retrieval-augmented verification (COIN-RAG) to ensure expert-level precision in analyzing raw, collective human discussions. An extensive evaluation of 20 state-of-the-art LLMs across four dimensions - depth, breadth, informativeness, and correctness - reveals that while current models can handle surface-level aggregation, they still struggle with the analytical depth required for complex intent synthesis. COIN-BENCH establishes a new standard for advancing LLMs from passive instruction followers to expert-level analytical agents capable of deciphering the collective voice of the real world. See our project page on COIN-BENCH.
title COINBench: Moving Beyond Individual Perspectives to Collective Intent Understanding
topic Information Retrieval
Artificial Intelligence
url https://arxiv.org/abs/2603.21329