Unverified50% confidenceOpinionExact time
When an LLM is asked to evaluate its own output, the scoring process shares the same model parameters and knowledge deficiencies as answer generation, creating circular dependence and failing to provide a verification signal independent of the answer.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIf the correct information has been fed into the LLM but the answer is still wrong, the problem may lie in how the model uses the context, such as instruction following, attention allocation, or output format constraints.79% similarUnverifiedHaving an LLM judge whether its own answers are correct is circular reasoning, because hallucinations and correct outputs may be structurally indistinguishable.77% similarUnverifiedLLM模型在「生成」和「评估」时所激活的推理路径不同，同一模型生成代码时可能忽略某个边界条件，但在评估另一份代码时却能识别该问题76% similarUnverifiedGeneration metrics usually use an LLM as the judge, computed with the help of libraries such as Ragas, and include faithfulness, answer relevance, and answer correctness.75% similarUnverifiedWhen automated scores fluctuate greatly but manual review finds the quality of multiple variant answers to be comparable, the problem lies in the evaluation stage rather than the model.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084178API
curl https://kongchang.com/api/v1/knowledge/claims/1084178MCP
get_claim(id=1084178)