UCE builds a typed, evolving library of Memory, Strategy, Workflow and Skill units from agent trajectories, improving ALFWorld success from 75.4% to 96.3% and WebShop score from 45.1% to 61.3% while transferring to new actor models.
From Exploration to Exploitation: A Two-Stage Entropy RLVR Approach for Noise-Tolerant MLLM Training.arXiv preprint arXiv:2511.07738, 2025
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
SGSD retrieves skill-mistake pairs to build a multi-teacher pool, validates teacher polarity via a verifier, and applies a gated objective to distill useful signals, yielding 6.2% average gains over GRPO on math benchmarks with Qwen3-1.7B.
citing papers explorer
-
Unified Context Evolution for LLM Agents
UCE builds a typed, evolving library of Memory, Strategy, Workflow and Skill units from agent trajectories, improving ALFWorld success from 75.4% to 96.3% and WebShop score from 45.1% to 61.3% while transferring to new actor models.
-
Skill-Conditioned Gated Self-Distillation for LLM Reasoning
SGSD retrieves skill-mistake pairs to build a multi-teacher pool, validates teacher polarity via a verifier, and applies a gated objective to distill useful signals, yielding 6.2% average gains over GRPO on math benchmarks with Qwen3-1.7B.