[KongchangAI]
Unverified50% confidenceFactExact time

The scorer produced two false-negative judgments on correct answers from Gemini 2.5 Flash because it only searched for specific phrasings such as 'zero', 'mask', 'approximate', and 'heuristic'.

1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027

Sources

Can LLMs Audit Machine Unlearning? A Benchmark That Exposes Fake Forgetting Code

Related Entities

Gemini 2.5 Flashmodel

Related Claims

UnverifiedConfirm-then-answer (probing before extracting within the same conversation) reduces each model's fabrication rate to 0–2%, while Gemini 3 Flash is also able to correctly calculate 98% of recoverable totals.64% similar

Cite This Claim

Stable URI
https://kongchang.com/claim/998825
API
curl https://kongchang.com/api/v1/knowledge/claims/998825
MCP
get_claim(id=998825)