Unverified50% confidenceBenchmarkExact time
After applying logit penalties, the number of reasoning tokens decreased by 11% to 19% across all precision formats.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/27/2026
First Seen
Valid until: 12/26/2026
Sources
Related Entities
Related Claims
UnverifiedApplying a logit bias of -2 reduces the probability of the target tokens to approximately 1/7 of the original (e^-2 ≈ 0.135).74% similarUnverifiedAfter removing the candidate filtering component, FEVER accuracy dropped to 76.24%.71% similarUnverifiedThe BF16 precision model's baseline accuracy was 74%, which improved to 84% after applying the penalty, with reasoning tokens reduced by 19.4%.70% similarUnverifiedWhen Token usage exceeds 60%–70%, the reasoning quality of Claude Code noticeably deteriorates.69% similarUnverified从BF16到FP8，内存占用减半，但数值表示精度大幅下降68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/954328API
curl https://kongchang.com/api/v1/knowledge/claims/954328MCP
get_claim(id=954328)