Unverified50% confidenceFactExact time
Previous research on LLM moral judgment has largely been based on two unrealistic assumptions: single-turn judgment and pressure-based rebuttal.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedResearch indicates that models may produce unfaithful chains of thought, where the surface-level reasoning steps are inconsistent with the actual basis for decision-making.72% similarUnverifiedThe evaluation of large models should go beyond the multiple-choice paradigm; testing only whether correct answers are selected is insufficient to measure genuine factual understanding, and adversarial evaluation that tests rejection of incorrect statements must be introduced.70% similarUnverifiedThe core hypothesis of the study is that external evidence is not simply accepted or rejected, but rather changes the probability distribution of the recipient's initial answer.70% similarUnverified直接让AI做逆向分析可能会触发模型的道德/合规约束而拒绝执行69% similarUnverifiedCounterfactual evaluation faces a fundamental problem of causal inference: it is impossible to simultaneously observe the outcomes for the same individual under both intervention and non-intervention conditions.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/898706API
curl https://kongchang.com/api/v1/knowledge/claims/898706MCP
get_claim(id=898706)