Expired50% confidenceFactExact time
Mainstream models such as GPT-4, Claude, and Gemini show significant deviation between confidence and accuracy on questions beyond their knowledge cutoff dates, multi-hop reasoning tasks, and long-tail knowledge domains.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/8/2026
First Seen
Valid until: 10/6/2026(expired)
Sources
元认知反馈：用强化学习让大模型准确表达不确定性
hackernewshackernews7/7/2026
Related Claims
Unverified纯大模型存在知识截止问题和幻觉问题两大天然缺陷，GPT-4的训练数据存在明确的截止日期82% similarUnverifiedMultimodal large models such as GPT-4, Claude 3.5, and Gemini can understand cross-file dependency relationships, infer architectural intent, and autonomously complete debugging when given error logs.77% similarUnverifiedMainstream large language models (including GPT-4, DeepSeek, Claude, etc.) are all trained on static corpora and have a definitive knowledge cutoff date.77% similarUnverifiedIn four rounds of testing by a Bilibili content creator, DeepSeek-V4 outperformed GPT-5.5 in delivery quality across three categories: world knowledge, contextual memory, and logical reasoning.74% similarUnverifiedBy 2024-2025, models like GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro achieved significant advances in logical reasoning, long-context understanding, and multimodal processing.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/261057API
curl https://kongchang.com/api/v1/knowledge/claims/261057MCP
get_claim(id=261057)