Jeffrey Ladish
Source-listed: Executive Director · Palisade Research
Record snapshot: Sep 3, 2026. Coverage and affiliations may be incomplete or historical.
Selected work
5- Language Models Can Autonomously Hack and Self-ReplicateSaved credit: Contributor · 2026Explore people connected to this work →
- Incomplete Tasks Induce Shutdown Resistance in Some Frontier LLMsSaved credit: Contributor · 2026Explore people connected to this work →
- Demonstrating specification gaming in reasoning modelsSaved credit: Contributor · 2025Explore people connected to this work →
- BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13BSaved credit: Contributor · 2023Explore people connected to this work →
- Constitutional AI: Harmlessness from AI FeedbackSaved credit: Contributor · 2022Explore people connected to this work →