LEAD uses online adaptive mechanisms including Potential-Scaled Instability and symmetric efficiency rewards based on correct rollouts to achieve higher accuracy-efficiency scores with substantially shorter reasoning outputs than base models on math benchmarks.
hub
Just enough thinking: Efficient reasoning with adaptive length penalties reinforcement learning
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
hub tools
citation-role summary
background 2
baseline 1
citation-polarity summary
verdicts
UNVERDICTED 13representative citing papers
DUET improves RLVR by allocating tokens across both prompt selection and rollout length, outperforming full-budget baselines even when using only half the tokens.
ThinkARM abstracts LLM reasoning traces into Schoenfeld episodes and shows that exploration steps correlate with correctness while efficiency methods selectively suppress evaluative feedback.
CLORE augments correct on-policy rollouts by deleting repetitive and irrelevant segments then optimizes with auxiliary DPO to improve accuracy-efficiency trade-off on math benchmarks.
CES applies conditional bidirectional entropy control on top of DAPO to improve accuracy and shorten responses on mathematical benchmarks for 7B and 1.5B LLMs.
ZAYA1-8B is a reasoning MoE model with 700M active parameters that matches larger models on math and coding benchmarks and reaches 91.9% on AIME'25 via Markovian RSA test-time compute.
LenVM models token-level remaining generation length as a bounded discounted value function derived from constant negative per-token rewards, providing a scalable proxy for generation horizon.
CODA uses rollout-based difficulty signals to drive two gates that penalize verbosity on easy instances and promote deliberation on hard ones, cutting token use over 60% on simple tasks while maintaining accuracy.
ATTNPO guides process-supervised RL with intrinsic attention signals to shorten reasoning traces while raising accuracy on nine benchmarks.
NCoTS treats chain-of-thought reasoning as a search problem and uses a dual-factor heuristic to find paths that are over 3.5% more accurate and 22% shorter on benchmarks.
TrOPD stabilizes on-policy distillation for LLMs with trust-region learning, outlier estimation, and off-policy guidance, outperforming prior OPD methods on reasoning and code benchmarks.
ZONOS2 8B is a scaled MoE TTS model with 900M active parameters trained on 6M hours of data that reports competitive SOTA results on naturalness, speaker similarity, WER, and a new ZTTS1-Eval benchmark while releasing weights and code.
Libra optimizes GPU allocation across rollout and training in agentic RL via an elastic hybrid pool and C-MLFQ scheduler based on tool-return causal signals, claiming up to 3.0x throughput and 2.5x faster reward convergence on 48 A800 GPUs.
citing papers explorer
-
LEAD: Length-Efficient Adaptive and Dynamic Reasoning for Large Language Models
LEAD uses online adaptive mechanisms including Potential-Scaled Instability and symmetric efficiency rewards based on correct rollouts to achieve higher accuracy-efficiency scores with substantially shorter reasoning outputs than base models on math benchmarks.
-
DUET: Optimize Token-Budget Allocation for Reinforcement Learning with Verifiable Rewards
DUET improves RLVR by allocating tokens across both prompt selection and rollout length, outperforming full-budget baselines even when using only half the tokens.
-
Schoenfeld's Anatomy of Mathematical Reasoning by Language Models
ThinkARM abstracts LLM reasoning traces into Schoenfeld episodes and shows that exploration steps correlate with correctness while efficiency methods selectively suppress evaluative feedback.
-
CLORE: Content-Level Optimization for Reasoning Efficiency
CLORE augments correct on-policy rollouts by deleting repetitive and irrelevant segments then optimizes with auxiliary DPO to improve accuracy-efficiency trade-off on math benchmarks.
-
Taming the Thinker: Conditional Entropy Shaping for Adaptive LLM Reasoning
CES applies conditional bidirectional entropy control on top of DAPO to improve accuracy and shorten responses on mathematical benchmarks for 7B and 1.5B LLMs.
-
ZAYA1-8B Technical Report
ZAYA1-8B is a reasoning MoE model with 700M active parameters that matches larger models on math and coding benchmarks and reaches 91.9% on AIME'25 via Markovian RSA test-time compute.
-
Length Value Model: Scalable Value Pretraining for Token-Level Length Modeling
LenVM models token-level remaining generation length as a bounded discounted value function derived from constant negative per-token rewards, providing a scalable proxy for generation horizon.
-
CODA: Difficulty-Aware Compute Allocation for Adaptive Reasoning
CODA uses rollout-based difficulty signals to drive two gates that penalize verbosity on easy instances and promote deliberation on hard ones, cutting token use over 60% on simple tasks while maintaining accuracy.
-
ATTNPO: Attention-Guided Process Supervision for Efficient Reasoning
ATTNPO guides process-supervised RL with intrinsic attention signals to shorten reasoning traces while raising accuracy on nine benchmarks.
-
Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Models
NCoTS treats chain-of-thought reasoning as a search problem and uses a dual-factor heuristic to find paths that are over 3.5% more accurate and 22% shorter on benchmarks.
-
Trust Region On-Policy Distillation
TrOPD stabilizes on-policy distillation for LLMs with trust-region learning, outlier estimation, and off-policy guidance, outperforming prior OPD methods on reasoning and code benchmarks.
-
ZONOS2 Technical Report
ZONOS2 8B is a scaled MoE TTS model with 900M active parameters trained on 6M hours of data that reports competitive SOTA results on naturalness, speaker similarity, WER, and a new ZTTS1-Eval benchmark while releasing weights and code.
-
Libra: Efficient Resource Management for Agentic RL Post-Training
Libra optimizes GPU allocation across rollout and training in agentic RL via an elastic hybrid pool and C-MLFQ scheduler based on tool-return causal signals, claiming up to 3.0x throughput and 2.5x faster reward convergence on 48 A800 GPUs.