Unverified50% confidenceFactExact time
Large models can hallucinate, and their outputs are not stable.
1
Sources
50%
Confidence
Long-term
Relevance
9/19/2026
First Seen
Sources
Related Entities
Related Claims
Unverified模型规模越大，幻觉的表述往往越流畅、越具迷惑性，识别难度反而上升79% similarUnverifiedPure large language models have two major deficiencies: a knowledge cutoff problem and a hallucination problem.77% similarVerifiedResearch indicates that hallucination rates do not have a linear negative correlation with model scale, and larger models sometimes produce harder-to-detect misinformation in a more fluent manner.76% similarUnverified大模型的幻觉问题目前只能缓解无法根除，知识库约束和精确模式是工程层面最有效的缓解手段73% similarUnverifiedPoor chunking will damage semantic features and increase the risk of hallucinations in the outputs generated by large models.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/935211API
curl https://kongchang.com/api/v1/knowledge/claims/935211MCP
get_claim(id=935211)