Emergent Attention Heads in Post-Training Reasoning Models
Discover how post-training sparks emergent attention heads that enhance reasoning models' performance and reveal key trade-offs in AI computation.
Mantis: Advanced Foundation Model for Disease Forecasting
Discover Mantis, a simulation-trained foundation model offering accurate, versatile disease forecasting even in low-data and novel outbreak settings.
Synthetic POMDPs for Memory-Augmented RL Challenges
Explore synthetic POMDPs to evaluate memory-augmented RL using Memory Demand Structure for scalable, tunable environment challenges.
League of LLMs: Benchmark-Free Evaluation for AI Models
Discover League of LLMs, a novel benchmark-free paradigm enabling mutual, transparent, and objective evaluation of large language models.
Order Effects and Biases in Large Language Models
Explore how order effects impact preferences in large language models and discover strategies to mitigate biases in decision-making systems.