AutoTool uses reinforcement learning with dual-mode rewards to train multimodal LLMs to adaptively choose between tool-assisted and text-centric reasoning, yielding accuracy and efficiency gains on V* and POPE benchmarks.
Title resolution pending
23 Pith papers cite this work. Polarity classification is still indexing.
23
Pith papers citing it
representative citing papers
StepCodeReasoner aligns code reasoning with verifiable stepwise execution traces via print anchors and bi-level GRPO reinforcement learning, reaching SOTA results on CRUXEval (91.1%) and LiveCodeBench (86.5%) for a 7B model.
SCOUT achieves state-of-the-art long-text understanding with up to 8x lower token use by actively foraging for sparse query-relevant information and updating a compact provenance-grounded epistemic state.
ReflectMT internalizes reflection via two-stage RL to enable direct high-quality machine translation that outperforms explicit reasoning models like DeepSeek-R1 on WMT24 while using 94% fewer tokens.
DelTA estimates token coefficients to amplify discriminative directions in token-gradient vectors, reweighting the RLVR surrogate to produce more contrastive side-wise centroids and yielding 3.26 and 2.62 point gains on math benchmarks for 8B and 14B Qwen3 models.
AutoVecCoder combines VecPrompt for automated intrinsic knowledge synthesis and VecRL for efficiency-aligned RL to train an 8B LLM that achieves SOTA on SimdBench SSE/AVX subsets and sometimes exceeds -O3 compiler results.
SAGE reshapes the reverse-KL anchor via guide function q(x,y) for controllable empirical support expansion, yielding gains in both pass@1 and pass@k on math reasoning benchmarks.
GRLO shows RLHF from scratch on 5K open-ended prompts raises average performance from 24.1 to 63.1 across domains on Qwen3-4B-Base using 46x less data and 68x less compute than in-domain RLVR while remaining competitive with heavily post-trained models.
A new RL method called MoCA with Perception Verification rewards perceptual fidelity independently to improve both seeing and thinking in VLMs.
ODRPO decomposes discrete rewards into ordinal binary indicators to create robust, variance-aware advantage estimators for noisy RLAIF in LLM alignment.
BalCapRL applies balanced multi-objective RL with GDPO-style normalization and length-conditional masking to improve MLLM image captioning, reporting gains of up to +13.6 DCScore, +9.0 CaptionQA, and +29.0 CapArena on LLaVA and Qwen models.
Listwise Policy Optimization explicitly performs target-projection on the LLM response simplex, unifying and improving group-based RLVR methods with monotonic improvement and flexible divergences.
BEACON uses milestone partitioning, temporal reward shaping, and dual-scale advantage estimation to nearly double success rates on long-horizon ALFWorld tasks while raising effective sample use from 23.7% to 82%.
Odysseus adapts PPO with a turn-level critic and leverages pretrained VLM action priors to train agents achieving at least 3x average game progress over frontier models in long-horizon Super Mario Land.
Waypoint-based bi-level planning with curriculum RLVR improves multi-robot task success rates in dense-obstacle benchmarks over motion-agnostic and VLA baselines.
Hybrid Policy Distillation unifies existing knowledge distillation methods for LLMs into a reweighted log-likelihood objective and introduces a hybrid forward-reverse KL approach with mixed data sampling to improve stability, efficiency, and performance.
HEAL mitigates entropy collapse in few-shot RLVR by selectively adding general-domain data and aligning trajectory-level entropy dynamics, matching full-shot performance with 32 target samples.
Recovering an orthogonal basis from model activations yields a model-native skill characterization that improves reasoning Pass@1 by up to 41% via targeted data selection and supports inference steering, outperforming human-characterized alternatives.
PRIME enables online process reward model updates in LLM RL using implicit rewards from rollouts and outcome labels, yielding 15.1% average gains on reasoning benchmarks and surpassing a stronger instruct model with 10% of the data.
TIF-GRPO uses integral feedback on pseudo-temporal trajectories to regulate anatomy-aware rewards in RL for clinical faithfulness in volumetric CT analysis.
Skill1 trains a single RL policy to co-evolve skill selection, utilization, and distillation in language model agents from one task-outcome reward, using low-frequency trends to credit selection and high-frequency variation to credit distillation, outperforming baselines on ALFWorld and WebShop.
Longer action horizons bottleneck LLM agent training through instability, but training with reduced horizons stabilizes learning and enables better generalization to longer horizons.
citing papers explorer
-
Are Tools Always Beneficial? Learning to Invoke Tools Adaptively for Dual-Mode Multimodal LLM Reasoning
AutoTool uses reinforcement learning with dual-mode rewards to train multimodal LLMs to adaptively choose between tool-assisted and text-centric reasoning, yielding accuracy and efficiency gains on V* and POPE benchmarks.
-
StepCodeReasoner: Aligning Code Reasoning with Stepwise Execution Traces via Reinforcement Learning
StepCodeReasoner aligns code reasoning with verifiable stepwise execution traces via print anchors and bi-level GRPO reinforcement learning, reaching SOTA results on CRUXEval (91.1%) and LiveCodeBench (86.5%) for a 7B model.
-
SCOUT: Active Information Foraging for Long-Text Understanding with Decoupled Epistemic States
SCOUT achieves state-of-the-art long-text understanding with up to 8x lower token use by actively foraging for sparse query-relevant information and updating a compact provenance-grounded epistemic state.
-
ReflectMT: Internalizing Reflection for Efficient and High-Quality Machine Translation
ReflectMT internalizes reflection via two-stage RL to enable direct high-quality machine translation that outperforms explicit reasoning models like DeepSeek-R1 on WMT24 while using 94% fewer tokens.
-
DelTA: Discriminative Token Credit Assignment for Reinforcement Learning from Verifiable Rewards
DelTA estimates token coefficients to amplify discriminative directions in token-gradient vectors, reweighting the RLVR surrogate to produce more contrastive side-wise centroids and yielding 3.26 and 2.62 point gains on math benchmarks for 8B and 14B Qwen3 models.
-
AutoVecCoder: Teaching LLMs to Generate Explicitly Vectorized Code
AutoVecCoder combines VecPrompt for automated intrinsic knowledge synthesis and VecRL for efficiency-aligned RL to train an 8B LLM that achieves SOTA on SimdBench SSE/AVX subsets and sometimes exceeds -O3 compiler results.
-
SAGE: Shaping Anchors for Guided Exploration in RLVR of LLMs
SAGE reshapes the reverse-KL anchor via guide function q(x,y) for controllable empirical support expansion, yielding gains in both pass@1 and pass@k on math reasoning benchmarks.
-
GRLO: Towards Generalizable Reinforcement Learning in Open-Ended Environments from Zero
GRLO shows RLHF from scratch on 5K open-ended prompts raises average performance from 24.1 to 63.1 across domains on Qwen3-4B-Base using 46x less data and 68x less compute than in-domain RLVR while remaining competitive with heavily post-trained models.
-
Bad Seeing or Bad Thinking? Rewarding Perception for Vision-Language Reasoning
A new RL method called MoCA with Perception Verification rewards perceptual fidelity independently to improve both seeing and thinking in VLMs.
-
ODRPO: Ordinal Decompositions of Discrete Rewards for Robust Policy Optimization
ODRPO decomposes discrete rewards into ordinal binary indicators to create robust, variance-aware advantage estimators for noisy RLAIF in LLM alignment.
-
BalCapRL: A Balanced Framework for RL-Based MLLM Image Captioning
BalCapRL applies balanced multi-objective RL with GDPO-style normalization and length-conditional masking to improve MLLM image captioning, reporting gains of up to +13.6 DCScore, +9.0 CaptionQA, and +29.0 CapArena on LLaVA and Qwen models.
-
Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex
Listwise Policy Optimization explicitly performs target-projection on the LLM response simplex, unifying and improving group-based RLVR methods with monotonic improvement and flexible divergences.
-
Milestone-Guided Policy Learning for Long-Horizon Language Agents
BEACON uses milestone partitioning, temporal reward shaping, and dual-scale advantage estimation to nearly double success rates on long-horizon ALFWorld tasks while raising effective sample use from 23.7% to 82%.
-
Odysseus: Scaling VLMs to 100+ Turn Decision-Making in Games via Reinforcement Learning
Odysseus adapts PPO with a turn-level critic and leverages pretrained VLM action priors to train agents achieving at least 3x average game progress over frontier models in long-horizon Super Mario Land.
-
Navigating the Clutter: Waypoint-Based Bi-Level Planning for Multi-Robot Systems
Waypoint-based bi-level planning with curriculum RLVR improves multi-robot task success rates in dense-obstacle benchmarks over motion-agnostic and VLA baselines.
-
Hybrid Policy Distillation for LLMs
Hybrid Policy Distillation unifies existing knowledge distillation methods for LLMs into a reweighted log-likelihood objective and introduces a hybrid forward-reverse KL approach with mixed data sampling to improve stability, efficiency, and performance.
-
HEALing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment
HEAL mitigates entropy collapse in few-shot RLVR by selectively adding general-domain data and aligning trajectory-level entropy dynamics, matching full-shot performance with 32 target samples.
-
Characterizing Model-Native Skills
Recovering an orthogonal basis from model activations yields a model-native skill characterization that improves reasoning Pass@1 by up to 41% via targeted data selection and supports inference steering, outperforming human-characterized alternatives.
-
Process Reinforcement through Implicit Rewards
PRIME enables online process reward model updates in LLM RL using implicit rewards from rollouts and outcome labels, yielding 15.1% average gains on reasoning benchmarks and surpassing a stronger instruct model with 10% of the data.
-
Regulating Anatomy-Aware Rewards via Trajectory-Integral Feedback for Volumetric Computed Tomography Analysis
TIF-GRPO uses integral feedback on pseudo-temporal trajectories to regulate anatomy-aware rewards in RL for clinical faithfulness in volumetric CT analysis.
-
Skill1: Unified Evolution of Skill-Augmented Agents via Reinforcement Learning
Skill1 trains a single RL policy to co-evolve skill selection, utilization, and distillation in language model agents from one task-outcome reward, using low-frequency trends to credit selection and high-frequency variation to credit distillation, outperforming baselines on ALFWorld and WebShop.
-
On Training Large Language Models for Long-Horizon Tasks: An Empirical Study of Horizon Length
Longer action horizons bottleneck LLM agent training through instability, but training with reduced horizons stabilizes learning and enables better generalization to longer horizons.
- SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows?