Mianqiu Huang

large language model (4) reinforcement learning from human feedback (2) preference alignment (1) in-context learning (1) confidence calibration (1) computational efficiency (1) safety alignment (1) probability calibration (1) model training (1) uncertainty estimation (1) memory optimization (1) context length (1) length extrapolation (1) inference-time alignment (1) long context (1) content safety (1) gpu memory (1) long-context modeling (1) safety training (1) dynamic alignment (1)