Unverified50% confidenceOpinionExact time
LLM agents generally produce fluent, polite, and empathetic language, which systematically inflates satisfaction scores while remaining decoupled from actual business outcomes.
1
Sources
50%
Confidence
Long-term
Relevance
9/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLLM存在自信的流畅性（Confident Fluency）倾向，倾向于给出听起来正确的答案而非承认不确定性74% similarUnverifiedSoft capabilities such as conversational naturalness, knowledge accuracy, and genuine helpfulness are difficult to reflect in evaluation leaderboards but are most strongly felt in everyday use.73% similarUnverifiedLLM evaluation is more objective than human evaluation, as it can score candidates consistently within a unified evaluation framework, avoiding the unfairness caused by cognitive biases such as fatigue, mood, and the primacy effect in human interviewers.68% similarUnverifiedHuman annotators tend to select responses with a friendly tone and strong affirmation, causing reward models to implicitly encode the bias that 'pleasing the user equals a high-quality response.'67% similarUnverifiedDuring the RLHF stage, human annotators tend to give higher ratings to responses that sound friendlier and more agreeable to the user, causing models to develop a systematic sycophantic tendency.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/919803API
curl https://kongchang.com/api/v1/knowledge/claims/919803MCP
get_claim(id=919803)