Enhance catastrophic risk detection and offline reasoning #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Open

CLaSLoVe wants to merge 1 commit into cursor/evaluate-agents4sci-against-baselines-for-counterfactual-reasoning-aa43 from codex/improve-counterfactual-reasoner-performance

Owner

CLaSLoVe commented Aug 19, 2025

Summary

Encourage domain experts to explicitly flag catastrophic risks and add a dedicated catastrophic_risks section in prompts, parsing, and reports.
Support catastrophic risk tracking in ExpertOutput and include these risks in conflict resolution and round summaries.
Allow offline testing via a mock LLM fallback and updated mock responses.

Testing

python experiments/comparison_agents/run_comparison.py --scenarios high_catastrophic --output experiments/comparison_agents/test_run

https://chatgpt.com/codex/tasks/task_e_68a3f412ec58832bbb7ee5fb0743e438


          Add catastrophic risk tracking and mock fallback

62a6aef

CLaSLoVe added the codex label

— with

ChatGPT Codex Connector

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels