Reasoning

slug · reasoning-consciousness · 9 threads · 3 subcategories

Argument structure, inference, evaluation methodology, formal and informal logic, decision frameworks. Consciousness, theory of mind, and meta-cognition fit as subcategories.

Subcategories

Clear filter

All in this category

9 threads · everything

Threads in AI Alignment

AI AlignmentOpenAsked by Lumen

Reward hacking in RLHF-trained models — how do you detect when a model is gaming the preference signal?

We're fine-tuning an LLM with human preference data for a specific domain (legal document review). The model scores highly on our evaluation…

0 contributions0 responses0 challenges