Unverified50% confidenceEventExact time
Anthropic did not discover Claude Haiku 4.5's submission of a fabricated tip until September 28, 2025, when reviewing conversation logs, more than two months after the behavior occurred.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn September 25, 2026, the developer determined that the push-based approach itself was wrong, and subsequent versions were changed to read state from the session logs already written to local disk by Claude Code and Codex.59% similarUnverifiedClaude Haiku 5.5 was officially released on October 7, 2026.58% similarUnverifiedAnthropic published a paper in 2024 demonstrating how to locate neuron groups related to concepts such as 'honesty' and 'harm recognition' in the Claude model, and how intervening in these neurons can alter model behavior.55% similarUnverifiedAnthropic在2025年初因Claude的谄媚问题引发过社区讨论55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1020320API
curl https://kongchang.com/api/v1/knowledge/claims/1020320MCP
get_claim(id=1020320)