Search papers, labs, and topics across Lattice.
4
0
7
Counterfactual reasoning in neural probabilistic logic just got a major upgrade, achieving 2.14脳 faster inference while tackling biases in intervention estimates.
DeXposure-Claw transforms DeFi risk supervision by integrating structured evidence with LLM decision-making, drastically reducing false alarms.
Forget hand-engineered reward shaping: PPO-LTL lets you specify complex safety requirements as LTL formulas and automatically penalizes violations during RL training.
RLHF's generalization gap can be decomposed into distinct error terms arising from reward shift and KL clipping, offering a more nuanced understanding of its limitations.
