Unverified50% confidenceBenchmarkExact time
2022年Chen等人在ACL发表的工作发现GPT系列模型在'解释为何好笑'任务上的表现（约62%人类满意度）显著低于'识别是否好笑'任务（约78%）
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
Why Can't AI Understand British Political Farce? The Cultural Comprehension Dilemma of Algorithms
hackernewshackernews7/8/2026
Related Claims
UnverifiedMultiple benchmark tests in 2023 (such as the HaHackathon dataset) showed that top LLMs still have significantly lower accuracy than ordinary human annotators on sarcasm recognition tasks66% similarUnverifiedA 2023 Stanford University study showed that multimodal models such as GPT-4V, when parsing flowcharts with nested conditional logic, had accuracy approximately 15–25 percentage points lower than equivalent textual descriptions.64% similarUnverifiedGPT 5.2 scored approximately 77% on competition problems in frontier science benchmarks, but its performance dropped to approximately 25% on open-ended research tasks.64% similarUnverified即便是GPT-4级别的Agent，面对含有5步以上工具调用的复杂任务时，端到端成功率往往低于60%61% similarUnverified2024年一项研究对GPT-4在Temperature=0下的代码生成任务进行100次测试，发现约8%的测试用例存在跨次运行的输出差异61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/411659API
curl https://kongchang.com/api/v1/knowledge/claims/411659MCP
get_claim(id=411659)