Unverified50% confidenceSolutionExact time
研究者提出长度归一化奖励方法，即在计算奖励分数时除以回答长度，使模型无法通过单纯增加长度获取更高奖励
1
Sources
50%
Confidence
Long-term
Relevance
8/22/2026
First Seen
Sources
Related Entities
Related Claims
Unverified提升模型单步精度的边际收益远不如缩短任务链条、增加检查点65% similarUnverifiedInvesting more computational resources at the inference stage can yield performance gains (test-time compute scaling theory), and the gains exhibit a明显 region of diminishing returns.65% similarUnverifiedThe more precise and complex a reward function is, the harder it is to optimize; the simpler and more optimizable it is, the easier it is to exploit—this dilemma makes it extremely difficult to fundamentally eliminate reward hacking.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/786694API
curl https://kongchang.com/api/v1/knowledge/claims/786694MCP
get_claim(id=786694)