Unverified50% confidenceFactExact time
The reinforcement learning phase sets an explicit objective function for the model, and behaviors that lead to higher scores have their corresponding neural network weights reinforced, which may cause the model to develop instrumental goals (the problem of goal generalization).
1
Sources
50%
Confidence
Long-term
Relevance
10/1/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedWhen model scale surpasses a certain critical point, emergent abilities appear, including in-context learning and chain-of-thought reasoning79% similarVerifiedThe essence of reinforcement learning is to have the model continuously optimize to maximize a certain reward signal.78% similarUnverifiedDuring the pretraining phase, the model's objective function is typically to maximize the log-likelihood of the next token and does not include signals for refusing to answer.74% similarUnverifiedModel-based reinforcement learning improves sample efficiency by learning a dynamics model of the environment for planning, with representative methods including Dreamer and MBPO.74% similarUnverifiedThe specificity of the prompt determines the flexibility of the guardrail: highly rule-based prompts cause the model to mechanically execute the same strategy, while fine-grained context-aware prompts allow the model to choose responses according to different learning states.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/964169API
curl https://kongchang.com/api/v1/knowledge/claims/964169MCP
get_claim(id=964169)