[KongchangAI]
Unverified50% confidenceFactExact time

The paper's core claim is that cross-lingual performance and social bias are two independent dimensions of LLM behavior, and English-centric fairness evaluation is insufficient to support models deployed globally.

1
Sources
50%
Confidence
Long-term
Relevance
11/10/2026
First Seen

Sources

Related Entities

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/1154792
API
curl https://kongchang.com/api/v1/knowledge/claims/1154792
MCP
get_claim(id=1154792)