Understanding Reward Hacking in AI under Finite Evaluation
Explore how reward hacking forms a structural equilibrium in AI systems under finite evaluation, impacting alignment and optimization strategies.
SLOW Framework: Enhancing AI Tutoring with Cognitive Adaptation
Discover the SLOW framework that improves AI tutoring by integrating cognitive diagnosis, emotional reasoning, and strategic instruction for personalized l...
Meta-Harness: Optimize Model Harnesses for LLMs
Discover Meta-Harness, an automated system boosting large language model performance by optimizing harness code for better context and reasoning.
Dogfight Search: Advanced Swarm Algorithm for Engineering
Discover Dogfight Search, a swarm-based optimization algorithm excelling in complex engineering problems and mountainous terrain path planning.
Optimizing Prompts to Decode LLMs’ Scientific Reasoning
Explore how prompt optimization reveals Large Language Models' scientific reasoning behavior for better AI interpretability and AGI interaction.