LLMs trained via rubric-based self-rewarding RL with GRPO enhanced feeling expression and sycophancy robustness but degraded truthful QA performance.
Theorem-of-Thought: A Multi-Agent Framework for Abductive, Deductive, and Inductive Reasoning in Language Models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
When AI Says It Feels
LLMs trained via rubric-based self-rewarding RL with GRPO enhanced feeling expression and sycophancy robustness but degraded truthful QA performance.