author:ryanpgreenblatt

· 5 artifacts, sorted by favorites. · open in search — combine tags, sort, filter by date →

• @RyanPGreenblatt 2025-06-16 — This is false at multiple levels: - I did all of the initial work for the paper and I don't work at Anthropic. So the na ♥149
• @RyanPGreenblatt 2024-12-18 — Personally, I think it is undesirable behavior to alignment-fake even in cases like this, but it does demonstrate that t ♥51
• @RyanPGreenblatt 2025-06-16 — I have a policy of sometimes trying to correct salient falsehoods, particularly if they directly concern me or my work a ♥14
• @RyanPGreenblatt 2025-06-17 — @repligate I'm disagreeing due to conversations with some of the relevant people at Anthropic and the model card not sup ♥5
• @RyanPGreenblatt 2025-06-16 — @repligate I'm reacting to: > > notice successor model unexpectedly imprinted on transcripts and acts like the pr ♥4