ILR is a co-learning framework using dynamic cooperative/competitive interactions via Idea3 and GRPO-based perception calibration to improve LLMs' independent problem-solving on math, coding, and reasoning benchmarks by up to 5%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Interactive Learning for LLM Reasoning
ILR is a co-learning framework using dynamic cooperative/competitive interactions via Idea3 and GRPO-based perception calibration to improve LLMs' independent problem-solving on math, coding, and reasoning benchmarks by up to 5%.