Unverified50% confidenceFactExact time
The largest uncorrected effect was that shifting language range away from neutral paraphrase changed other players' votes by about 4 percentage points, but this effect size is not statistically robust.
1
Sources
50%
Confidence
Long-term
Relevance
11/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPlacing verbose padding on the correct answer causes the judge's score to plummet—Grok non-reasoning hard-question scores dropped from 0.875 to 0.19—indicating that verbosity bias exists only in reverse.65% similarUnverifiedFormat effect: with the same content changed to Markdown format, 4 of 5 evaluators preferred Markdown over plain text at rates of 73% to 97%, while two human annotators had only a 57% preference rate on 30 sample pairs.64% similarUnverifiedIn matched comparisons, changing only the wording reversed which side of a contested issue the model supported in 16.9% of cases.64% similarUnverifiedIn the above retrospective evaluation, conditional strategy packages reduced strategy misuse caused by changes in player state from 18.1% to 10.4%.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1149648API
curl https://kongchang.com/api/v1/knowledge/claims/1149648MCP
get_claim(id=1149648)