Unverified50% confidenceFactExact time
The pre-registered hypothesis H2 (that more thinking is harmful on the distractor counting task) was not significantly supported, with no model performing significantly worse at high.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe authors state they will not provide unverified rankings and do not claim that any model comprehensively outperforms others, because a perfect score on 72 fixed cases cannot prove that a model performs equally well on unseen samples.67% similarUnverified对比学习中的推远操作只要求负样本相似度低于正样本，并不要求达到-1的极端值66% similarUnverifiedThe binary threshold design is unfavorable to models near the threshold, as 0.74 and 0.80 differ drastically in score even though the actual capability gap may be negligible.66% similarUnverifiedScoring higher than a previous-generation model on metrics does not equate to adequately resolving alignment issues.65% similarUnverified模型的基准分数更高并不等于在所有场景下都更好用64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1007698API
curl https://kongchang.com/api/v1/knowledge/claims/1007698MCP
get_claim(id=1007698)