LBR performs token-level test-time scaling via local branch routing on hidden states, enabling end-to-end RL training and improving Pass@1 and Pass@32 on math benchmarks over CoT and RLVR baselines.
Llms are single-threaded reasoners: Demystifying the working mechanism of soft thinking.arXiv preprint arXiv:2508.03440
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
years
2026 7representative citing papers
CopT reverses CoT by eliciting a draft answer first then using continuous-embedding contrastive verification and on-policy thinking to reflect and correct, yielding up to 23% higher accuracy and 57% fewer tokens without training.
Latent reasoning faithfulness is a property of training stage and answer format, not of architecture or the final checkpoint alone.
SOLAR aligns soft-token probability mixtures across languages in embedding space during SFT and raises multilingual reasoning accuracy by up to 17.7 points over the base model.
SeLaR selectively applies latent soft reasoning in LLMs via entropy gating and contrastive regularization, outperforming standard CoT on five benchmarks without training.
TARPO is a pure RL framework using a token-wise action router to switch between discrete token generation and latent reasoning in LLMs, with joint optimization showing outperformance on benchmarks.
LEPO applies RL to continuous latent representations in LLMs by injecting Gumbel-Softmax stochasticity for diverse trajectory sampling and unified gradient estimation, outperforming existing discrete and latent RL methods.
citing papers explorer
-
Efficient and Trainable Language Model Test-Time Scaling via Local Branch Routing
LBR performs token-level test-time scaling via local branch routing on hidden states, enabling end-to-end RL training and improving Pass@1 and Pass@32 on math benchmarks over CoT and RLVR baselines.
-
CopT: Contrastive On-Policy Thinking with Continuous Spaces for General and Agentic Reasoning
CopT reverses CoT by eliciting a draft answer first then using continuous-embedding contrastive verification and on-policy thinking to reflect and correct, yielding up to 23% higher accuracy and 57% fewer tokens without training.
-
Final Checkpoints Are Not Enough: Analyzing Latent Reasoning Faithfulness Along Training Trajectories
Latent reasoning faithfulness is a property of training stage and answer format, not of architecture or the final checkpoint alone.
-
Soft Token Alignment for Cross-Lingual Reasoning
SOLAR aligns soft-token probability mixtures across languages in embedding space during SFT and raises multilingual reasoning accuracy by up to 17.7 points over the base model.
-
SeLaR: Selective Latent Reasoning in Large Language Models
SeLaR selectively applies latent soft reasoning in LLMs via entropy gating and contrastive regularization, outperforming standard CoT on five benchmarks without training.
-
TARPO: Token-Wise Latent-Explicit Reasoning via Action-Routing Policy Optimization
TARPO is a pure RL framework using a token-wise action router to switch between discrete token generation and latent reasoning in LLMs, with joint optimization showing outperformance on benchmarks.
-
LEPO: Latent Reasoning Policy Optimization for Large Language Models
LEPO applies RL to continuous latent representations in LLMs by injecting Gumbel-Softmax stochasticity for diverse trajectory sampling and unified gradient estimation, outperforming existing discrete and latent RL methods.