SPOT: Sparse Probing and Outcome Calibration for On-Policy Distillation
Paper • 2608.04419 • Published • 11
Zhongxiang Dai
dzxagent
AI & ML interests
Agents, LLMs
Recent Activity
authored a paper about 2 hours ago
SPOT: Sparse Probing and Outcome Calibration for On-Policy Distillation authored a paper about 2 hours ago
DRIFT: Decoupled Rollouts and Importance-Weighted Fine-Tuning for Efficient Multi-Turn Optimization authored a paper about 2 hours ago
MASPOB: Bandit-Based Prompt Optimization for Multi-Agent Systems with Graph Neural NetworksOrganizations
None yet
authored 16
papers about 2 hours ago
DRIFT: Decoupled Rollouts and Importance-Weighted Fine-Tuning for Efficient Multi-Turn Optimization
Paper • 2605.31455 • Published • 6
MASPOB: Bandit-Based Prompt Optimization for Multi-Agent Systems with Graph Neural Networks
Paper • 2603.02630 • Published
Words & Weights: Streamlining Multi-Turn Interactions via Co-Adaptation
Paper • 2603.01375 • Published • 1
T-POP: Test-Time Personalization with Online Preference Feedback
Paper • 2509.24696 • Published • 1
Adaptive Batch-Wise Sample Scheduling for Direct Preference Optimization
Paper • 2506.17252 • Published • 2
Real-Time Aligned Reward Model beyond Semantics
Paper • 2601.22664 • Published • 15
Self-Reflective Generation at Test Time
Paper • 2510.02919 • Published • 10
Federated Zeroth-Order Optimization using Trajectory-Informed Surrogate Gradients
Paper • 2308.04077 • Published
Data-Centric AI in the Age of Large Language Models
Paper • 2406.14473 • Published • 1
Prompt Optimization with EASE? Efficient Ordering-aware Automated Selection of Exemplars
Paper • 2405.16122 • Published
Localized Zeroth-Order Prompt Optimization
Paper • 2403.02993 • Published
Use Your INSTINCT: INSTruction optimization for LLMs usIng Neural bandits Coupled with Transformers
Paper • 2310.02905 • Published
Prompt Optimization with Human Feedback
Paper • 2405.17346 • Published
Robustifying and Boosting Training-Free Neural Architecture Search
Paper • 2403.07591 • Published
Training-Free Neural Active Learning with Initialization-Robustness Guarantees
Paper • 2306.04454 • Published
upvoted a paper about 7 hours ago