Unverified50% confidenceFactExact time
Models were not explicitly trained to persuade, yet they naturally acquired persuasion capabilities through language generation; researchers call this emergent persuasion ability.
1
Sources
50%
Confidence
Long-term
Relevance
9/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe pretraining corpora of the vast majority of models are closed-source, forcing membership inference research to rely on heuristic methods to construct members and non-members, which may introduce systematic bias.75% similarUnverifiedThe root cause of rising hallucination rates in reasoning models is misaligned incentives: reinforcement learning training causes models to tend to generate detailed and persuasive chains of thought, and human evaluators give higher scores to more elaborately argued answers, even when they contain factual errors.72% similarUnverifiedDirectly importing raw chat logs into a fine-tuning dataset yields poor training results, because raw chat logs lack emotional and contextual annotations, leaving the model only able to learn surface character patterns rather than the decision logic of when to use a certain style.72% similarUnverifiedWhen early model capabilities were insufficient, large amounts of guiding prompts were needed, but those prompts in turn constrained the boundaries of the model's capabilities.71% similarUnverifiedThis method is the simplest form of imitation learning—behavioral cloning—bypassing the complex reward design and exploration problems in reinforcement learning.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/932243API
curl https://kongchang.com/api/v1/knowledge/claims/932243MCP
get_claim(id=932243)