跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.25459 · v1

PILAPS: A Bilingual Benchmark for Evaluating Pragmatic Inference across Phenomena and Settings

首次提交 2026-11-19 11:44
版本更新 2026-11-19 11:44
雷达观测 2026-11-20 10:50
arXiv ↗PDF ↗

摘要

Recent work reports strong pragmatic performance by large language models (LLMs) in English, yet whether this performance generalises across languages and evaluation settings remains unclear. This question is increasingly important as LLMs are deployed across languages, cultures, and conversational contexts. In this paper, we first introduce PILAPS, a bilingual English/Mandarin Chinese benchmark covering eight pragmatic phenomena across purpose-written and dialogue-derived items, with both culturally adapted and independently sourced data. We then evaluate eight LLMs from four families across a range of model sizes and prompting strategies. Our results reveal substantial variation across models and pragmatic phenomena: open-source models continue to struggle particularly with humour, while cross-linguistic performance varies with model origin. Our error analysis shows that, compared with human participants, models are more prone to literal interpretations in conversational contexts. Theory-of-Mind prompting lowers the literal share of errors, whereas Chain-of-Thought prompting slightly raises it. Together, these findings demonstrate that strong pragmatic performance in one evaluation setting does not necessarily generalise across languages or conversational contexts.

作者备注:23 pages, 4 figures, 19 tables

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-19 11:44 · 最新观测版本

PILAPS: A Bilingual Benchmark for Evaluating Pragmatic Inference across Phenomena and Settings

Recent work reports strong pragmatic performance by large language models (LLMs) in English, yet whether this performance generalises across languages and evaluation settings remains unclear. This question is increasingly important as LLMs are deployed across languages, cultures, and conversational contexts. In this paper, we first introduce PILAPS, a bilingual English/Mandarin Chinese benchmark covering eight pragmatic phenomena across purpose-written and dialogue-derived items, with both culturally adapted and independently sourced data. We then evaluate eight LLMs from four families across a range of model sizes and prompting strategies. Our results reveal substantial variation across models and pragmatic phenomena: open-source models continue to struggle particularly with humour, while cross-linguistic performance varies with model origin. Our error analysis shows that, compared with human participants, models are more prone to literal interpretations in conversational contexts. Theory-of-Mind prompting lowers the literal share of errors, whereas Chain-of-Thought prompting slightly raises it. Together, these findings demonstrate that strong pragmatic performance in one evaluation setting does not necessarily generalise across languages or conversational contexts.

查看此版本解读 →查看此版本原文 ↗