Enhancing Physics Problem Solving with Retrieval-Augmented Models
Discover how retrieval-augmented generation boosts foundation models in solving Olympiad-level physics problems using the PhoPile dataset.
SeedPrints: Trace Your LLM’s Training Seed Easily
Discover SeedPrints, a breakthrough method to fingerprint large language models from their initial training seed for reliable model provenance.
HiCoLoRA: Advanced Zero-Shot Dialog State Tracking
Discover HiCoLoRA, a novel framework solving context-prompt misalignment for zero-shot dialog state tracking in task-oriented systems.
DyBBT: Adaptive Dialog Policy with Bandit-Inspired Dual Systems
Discover DyBBT, a dynamic dialog policy framework using bandit-inspired targeting and cognitive dual-systems for improved AI conversation efficiency.
Safe-SAIL: Fine-Grained Safety Analysis of Large Language Models
Discover Safe-SAIL, a framework using sparse autoencoders for detailed safety analysis and interpretability of large language models in critical domains.