Unverified50% confidenceFactExact time
When all sampled outputs in a group are incorrect, the group-relative advantage collapses to zero, causing the model to receive no effective gradient feedback in high-uncertainty scenarios.
1
Sources
50%
Confidence
Long-term
Relevance
9/25/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMisrouting sends complex requests to small models with insufficient capability, producing answers that appear plausible but are actually incorrect, and the cost of this quality degradation is difficult to quantify.76% similarUnverifiedOne of the main risks of implementing inference from scratch is numerical drift, where precision issues or incorrect computation order at any step may be amplified after propagating through multiple layers.75% similarUnverified如果参数量充足的模型连几十个样本都无法记住，几乎可以断定存在硬性错误，如标签错位、损失函数计算错误、梯度未正确回传等75% similarVerifiedThe double descent phenomenon indicates that when the number of model parameters far exceeds the amount of data, the test error may actually decrease again.74% similarUnverifiedThe margin of error only reflects the uncertainty introduced by sampling and cannot encompass non-sampling errors, such as question wording bias, non-response bias, coverage bias, and social desirability bias.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/949034API
curl https://kongchang.com/api/v1/knowledge/claims/949034MCP
get_claim(id=949034)