Sponsor of the Day:
Jerkmate
https://palisaderesearch.org/blog/badllama
Badllama: cheaply removing safety fine-tuning from Llama 2-Chat 13B | Palisade Research
Oct 31, 2023 - Llama 2-Chat is a collection of large language models that Meta developed and released to the public. While Meta fine-tuned Llama 2-Chat to refuse to output...
fine tuningllama 2palisade researchcheaplyremoving
https://palisaderesearch.org/
Home | Palisade Research
Apr 15, 2026 - AI capabilities are improving rapidly. We study the capabilities and motivations of AI agents today to better understand the risk of losing control to AI...
palisade research
https://palisaderesearch.org/about
About | Palisade Research
Apr 15, 2026 - Palisade Research is a nonprofit investigating cyber offensive AI capabilities and the controllability of frontier AI models. Our research has been highlighted...
palisade research
https://palisaderesearch.org/blog/gpt5-at-ctfs
GPT-5 at CTFs: case studies from top cybersecurity events | Palisade Research
Nov 20, 2025 - OpenAI and DeepMind’s AIs recently got gold at the IMO math olympiad and ICPC programming competition. We show frontier AI is similarly good at hacking by...
gpt 5case studiestop cybersecuritypalisade researchctfs