Token-Level Prompt Optimization for Diffusion Models
Discover how evolutionary token-level prompt optimization enhances text-to-image diffusion models for better image quality and alignment.
Teaching LLMs to Negotiate via Reinforcement Learning
Learn how Reinforcement Learning with verifiable rewards trains LLMs to master negotiation, outperforming larger models in strategic bargaining.
MEMENTO: Boost LLMs Context Management & Efficiency
Discover MEMENTO, a method that improves LLMs by managing context efficiently, reducing cache use, and enhancing reasoning accuracy across tasks.
Non-Surjective Nature of Steered LLM Activations
Discover why steered LLM activations are non-surjective, limiting prompt-based control and impacting AI interpretability and safety research.
COMPOSITE-STEM: Benchmarking AI for Scientific Discovery
Discover COMPOSITE-STEM, a new AI benchmark with 70 expert tasks across STEM fields, enhancing AI evaluation in scientific discovery workflows.