Post-Reasoning boosts LLM accuracy by reversing the usual answer-after-reasoning order, delivering mean relative gains of 17.37% across 117 model-benchmark pairs with zero extra cost.
Mirror descent policy optimization
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
representative citing papers
LPO reframes group-based RLVR as explicit target-projection on the LLM response simplex and performs exact divergence minimization to achieve monotonic listwise improvement with bounded gradients.
Muon optimizer with weight decay and update scaling achieves ~2x efficiency over AdamW for large LLMs, shown via the Moonlight 3B/16B MoE model trained on 5.7T tokens.
A new posterior sampling algorithm for (ε, δ)-PAC policy identification in tabular MDPs achieves asymptotic optimality in sample complexity and posterior contraction rate with O(S²AH) runtime per episode.
A survey organizing techniques to achieve efficient reasoning in LLMs by shortening chain-of-thought outputs.
citing papers explorer
-
Post Reasoning: Improving the Performance of Non-Thinking Models at No Cost
Post-Reasoning boosts LLM accuracy by reversing the usual answer-after-reasoning order, delivering mean relative gains of 17.37% across 117 model-benchmark pairs with zero extra cost.
-
Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex
LPO reframes group-based RLVR as explicit target-projection on the LLM response simplex and performs exact divergence minimization to achieve monotonic listwise improvement with bounded gradients.
-
Muon is Scalable for LLM Training
Muon optimizer with weight decay and update scaling achieves ~2x efficiency over AdamW for large LLMs, shown via the Moonlight 3B/16B MoE model trained on 5.7T tokens.
-
Optimal Posterior Sampling for Policy Identification in Tabular Markov Decision Processes
A new posterior sampling algorithm for (ε, δ)-PAC policy identification in tabular MDPs achieves asymptotic optimality in sample complexity and posterior contraction rate with O(S²AH) runtime per episode.
-
Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models
A survey organizing techniques to achieve efficient reasoning in LLMs by shortening chain-of-thought outputs.