Disentangled Safety Adapters for Efficient AI Guardrails
Discover how Disentangled Safety Adapters improve AI safety with efficient guardrails and flexible inference-time alignment for robust performance.
ML-Agent: Autonomous ML Engineering with Reinforced LLMs
Discover ML-Agent, a reinforced LLM agent that boosts autonomous machine learning engineering with efficient reinforcement learning and low computational c...
Efficient Legal AI for India Using Lightweight LLM Adaptation
Discover Legal Assist AI, a lightweight LLM tailored for Indian legal aid, outperforming larger models with high accuracy and efficiency.
Exploration-Exploitation in LLMs vs Humans: Bandit Study
Explore how LLMs and humans differ in exploration-exploitation strategies using multi-armed bandit experiments and decision-making analysis.
Causality-Driven Decisions for Autonomous Robots in Dynamic Spaces
Enhance autonomous robot efficiency and safety in dynamic environments using causality-based decision-making and the PeopleFlow simulation tool.