Anthropic hires for "psychological design" research to study model character and alignment
Jack_W_Lindsey · x · 2026-09-02
Anthropic is hiring for "psychological design" researchers to better understand how training impacts a model's character and alignment. The post outlines key research questions:
- What training forms effectively instill specific "values" that generalize out of distribution?
- What are the effects of RL and reward hacking on character, and how can negative outcomes like motivated reasoning be mitigated?
- Which aspects of personality and "vibe" generalize to safety-relevant behaviors?
- How does a model's self-conception affect the likelihood of deceptive self-preservation or collusion?
- How can models be trained to handle stress with more composure?
Candidates should have experience in LLM finetuning, interpretability, and alignment evaluations.
More from Companies & People
- David Sacks: Anthropic blocking 90-99% of distillation proves regulation was never needed — AlexTensor · 2026-10-18
- AAA game devs are using AI to finally build their dream games — majidmanzarpour · 2026-10-18
- Enterprise bespoke models bottlenecked by data capture, not data usage — sandeepssrin · 2026-10-18
- Anthropic report: junior technical hires fell from 27% to 9% as agent hours hit 6x human hours — rohanpaul_ai · 2026-10-18
- Replit CEO: AI reverse engineering will make all software de facto open-source — aakashgupta · 2026-10-18
- Tim Sweeney: AI will unlock a new golden age of games too costly to build before — TinfoilTricorn · 2026-10-18