Found 871 bookmarks
Custom sorting
OpenEQA: From word models to world models
OpenEQA: From word models to world models
OpenEQA combines challenging open-vocabulary questions with the ability to answer in natural language. This results in a straightforward benchmark that demonstrates a strong understanding of the environment—and poses a considerable challenge to current foundational models. We hope this work motivates additional research into helping AI understand and communicate about the world it sees.
·ai.meta.com·
OpenEQA: From word models to world models
Anil, C., Durmus, E., Sharma, M., Benton, J., Kundu, S., Batson, J., ... & Duvenaud, D. (2024). Many-shot Jailbreaking.
Anil, C., Durmus, E., Sharma, M., Benton, J., Kundu, S., Batson, J., ... & Duvenaud, D. (2024). Many-shot Jailbreaking.

Long contexts represent a new front in the struggle to control LLMs. We explored a family of attacks that are newly feasible due to longer context lengths, as well as candidate mitigations. We found that the effectiveness of attacks, and of in-context learning more generally, could be characterized by simple power laws. This provides a richer source of feedback for mitigating long-context attacks than the standard approach of measuring frequency of success

·www-cdn.anthropic.com·
Anil, C., Durmus, E., Sharma, M., Benton, J., Kundu, S., Batson, J., ... & Duvenaud, D. (2024). Many-shot Jailbreaking.