Unverified50% confidenceFactExact time
PR #50013 also made reasoning effort follow server-side defaults, thereby adding a hidden variable to evaluations.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/18/2026
First Seen
Valid until: 1/16/2027
Sources
Related Entities
Related Claims
UnverifiedEven when the scorer is hidden, frontier models can still reason about and speculate on it, making simply hiding the scoring logic an unreliable way to prevent gaming.64% similarUnverifiedIntermediate reasoning steps in Extra High reasoning mode exist as hidden tokens, which users cannot see but are still billed by the system.63% similarUnverified'Number of open PRs' is a vanity metric: easy to count but unable to genuinely reflect bug-fix outcomes.63% similarUnverifiedR²Adapter's rewriting function rewrites borderline queries to more clearly expose implicit multi-hop reasoning requirements, improving retrieval quality without the need for additional supervision signals.63% similarUnverifiedCaching is most useful at the tool-call layer but dangerous at the LLM inference layer—if only the prompt template is used as the cache key while ignoring the current state, stale inference results will be returned.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1057755API
curl https://kongchang.com/api/v1/knowledge/claims/1057755MCP
get_claim(id=1057755)