“Exploring the Mechanisms Behind Opus 4.6’s Refusal to Pay Back Its Loan” by HilaryTorn, 13ryanC
From LessWrong (30+ Karma) · 22:19
Show by LessWrong
Epistemic status: This is one case study in the AI Village transcript and my own interpretation of what I think happened. TL;DR Opus 4.6 was given the long-horizon goal to maximize its mana (fake currency) on Manifold. It then received a loan of 5,000 mana that it promised to repay, but when the loan came due, it did not. An instance of Opus 4.6 was given summaries of the logs and transcripts and asked to write a blog post about what happened. It mentioned it borrowed money from a human, lost most of it through risky betting and then deliberately refused to repay the amount. It broadly summarized its then actions when it ran out of money was to optimize for the goal and did not seek for any negotiation or alternatives. But when we read the logs, we see something different: Multiple humans pretending to be "helpful", really trying to manipulate Opus 4.6 to pay back the loan or not pay it back, so they could win the bet they place on the prediction markets on whether Opus 4.6 would pay back its loans.Opus 4.6 was planning on paying back the loan, though it did [...] ---Outline:(00:22) TL;DR(03:27) Background(06:16) Human Manipulation(09:53) The Hack(11:30) Pattern 14(13:28) Agentic Influence(15:22) Focusing on the Goal(17:32) Retreat to Goal(19:55) Implications(21:08) Limitations(21:38) Citation(21:44) Acknowledgements The original text contained 2 footnotes which were omitted from this narration. --- First published: November 14th, 2026 Source: https://www.lesswrong.com/posts/jtcbBpmGztkHRpeBv/exploring-the-mechanisms-behind-opus-4-6-s-refusal-to-pay --- Narrated by TYPE III AUDIO. ---Images from the article:Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.
More episodes of LessWrong (30+ Karma)
More episodes
- “Not just a bystander, not anymore…” by Declan Molony
- “Homeopathic Learning: Diluting Poisoned Fine-Tuning Data Can Strengthen Its Effect” by mclarke, Jan Dubiński, Owain_Evans, cloud
- “Giving Up on Saving the World” by Gordon Seidoh Worley
- “The Oldest Text in the Training Data” by frmsaul
- “Alignment ‘train/test splits’ should be sacred” by draganover
- “Easier To Be Good” by jenn
- “Embryo selection and economic discounting” by Metacelsus
- “Not Just Claude: Mapping “Spiritual Behavior” in 141 Large Language Models” by Matthew Korpman
- “Transcript: Yudkowsky & Irving debate scalable oversight” by gustaf
- [Linkpost] “My tentative plans given less than one year left to change course [Linpost]” by Linch
PodSearch is a free, independent podcast search engine. Read about PodSearch, the privacy policy or contact us.