Unverified50% confidenceFactExact time
GRPO and related policy gradient methods compress a complete multi-turn interaction trajectory into a scalar reward before feeding it into the policy update, which causes lossy credit assignment when handling composite skill tasks.
1
Sources
50%
Confidence
Long-term
Relevance
9/24/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe grep loop skill has the review Agent assign a confidence score (1–5) to the PR and leave feedback; a score below the threshold triggers a rebuild loop, and only after receiving a 5/5 does a human appear to click merge.74% similarUnverified基于策略梯度的算法如PPO和GRPO能够处理离散与连续混合的动作空间69% similarUnverifiedThe single-rollout characteristic of SAO means that a more streamlined sampling and trajectory generation mechanism is used at each policy update step, reducing reliance on multiple rollouts.69% similarVerifiedPPO (Proximal Policy Optimization) limits the step size of each policy update through a clipped objective function, and has become the de facto standard for continuous control tasks.65% similarUnverified模型在PPO优化中通过增加输出长度获取更高奖励的现象被称为reward hacking（奖励攻击）64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/946116API
curl https://kongchang.com/api/v1/knowledge/claims/946116MCP
get_claim(id=946116)