AcquisitionSynthesis uses acquisition functions as rewards to train generators that produce higher-quality synthetic data, delivering 2-7% gains on math, medical QA, and coding tasks with improved robustness to forgetting.
Title resolution pending
19 Pith papers cite this work. Polarity classification is still indexing.
19
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
Prepending stochastic sequences from Lorem Ipsum vocabulary to prompts during GRPO resampling broadens reasoning exploration and outperforms standard resampling on hard tasks for 1.7B-7B models.
UCPO modifies GRPO with a uniformity penalty over correct solutions to prevent diversity collapse in RLVR, yielding up to 10% higher Pass@64 on AIME24 and 45% more equation-level diversity.
Short GRPO warm-up followed by offline DPO on informative rollouts matches or beats full GRPO on math reasoning benchmarks at substantially lower compute cost.
GAGPO computes step-aligned temporal advantages from grouped rollout samples without a learned critic, enabling stable policy optimization in multi-turn agent environments.
Sync-R1 applies cooperative RL with Sync-GRPO and Dynamic Group Scaling to achieve superior cross-task personalized reasoning in multimodal models on the new UnifyBench++ dataset.
VIGOR assigns higher rewards to LLM completions that produce smaller l2 norms of teacher-forced negative log-likelihood gradients, with sqrt(T) length correction and group ranking, yielding +3.31% math and +1.91% code gains over RLIF on Qwen2.5-7B.
OPT-BENCH trains LLMs on NP-hard optimization via quality-aware RLVR, achieving 93.1% success rate and 46.6% quality ratio on Qwen2.5-7B while outperforming GPT-4o and transferring gains to other domains.
RPSFT improves the in-domain versus out-of-domain performance trade-off during LLM supervised fine-tuning by penalizing rotations in pretrained singular subspaces as a proxy for loss-sensitive directions.
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.
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%.
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.
Math reasoning gains in LLMs rarely transfer to general domains; RL tuning generalizes while SFT causes forgetting and representation drift.
Defines Entropy-Gradient Inversion as a geometric fingerprint of LRM reasoning and introduces CorR-PO to embed it in RL reward regularization, reporting improved benchmark performance.
DyGRO-VLA is a two-stage optimization framework for cross-task scaling of Vision-Language-Action models via dynamic grouped residual optimization in RL.
On-policy distillation gains efficiency from early foresight in module allocation and update directions, which the proposed EffOPD method exploits for 3x faster training with comparable performance.
DACA-GRPO adds denoising-aware credit assignment and bias-reduced likelihood estimation to GRPO, delivering consistent gains up to 36.3pp on math, code, constraint, and schema benchmarks for diffusion LLMs.
citing papers explorer
-
AcquisitionSynthesis: Targeted Data Generation using Acquisition Functions
AcquisitionSynthesis uses acquisition functions as rewards to train generators that produce higher-quality synthetic data, delivering 2-7% gains on math, medical QA, and coding tasks with improved robustness to forgetting.
-
Nonsense Helps: Prompt Space Perturbation Broadens Reasoning Exploration
Prepending stochastic sequences from Lorem Ipsum vocabulary to prompts during GRPO resampling broadens reasoning exploration and outperforms standard resampling on hard tasks for 1.7B-7B models.
-
Uniform-Correct Policy Optimization: Breaking RLVR's Indifference to Diversity
UCPO modifies GRPO with a uniformity penalty over correct solutions to prevent diversity collapse in RLVR, yielding up to 10% higher Pass@64 on AIME24 and 45% more equation-level diversity.
-
How Much Online RL is Enough? Informative Rollouts for Offline Preference Optimization in RLVR
Short GRPO warm-up followed by offline DPO on informative rollouts matches or beats full GRPO on math reasoning benchmarks at substantially lower compute cost.
-
GAGPO: Generalized Advantage Grouped Policy Optimization
GAGPO computes step-aligned temporal advantages from grouped rollout samples without a learned critic, enabling stable policy optimization in multi-turn agent environments.
-
Uni-Synergy: Bridging Understanding and Generation for Personalized Reasoning via Co-operative Reinforcement Learning
Sync-R1 applies cooperative RL with Sync-GRPO and Dynamic Group Scaling to achieve superior cross-task personalized reasoning in multimodal models on the new UnifyBench++ dataset.
-
Verifier-Free RL for LLMs via Intrinsic Gradient-Norm Reward
VIGOR assigns higher rewards to LLM completions that produce smaller l2 norms of teacher-forced negative log-likelihood gradients, with sqrt(T) length correction and group ranking, yielding +3.31% math and +1.91% code gains over RLIF on Qwen2.5-7B.
-
Forge: Quality-Aware Reinforcement Learning for NP-Hard Optimization in LLMs
OPT-BENCH trains LLMs on NP-hard optimization via quality-aware RLVR, achieving 93.1% success rate and 46.6% quality ratio on Qwen2.5-7B while outperforming GPT-4o and transferring gains to other domains.
-
Rotation-Preserving Supervised Fine-Tuning
RPSFT improves the in-domain versus out-of-domain performance trade-off during LLM supervised fine-tuning by penalizing rotations in pretrained singular subspaces as a proxy for loss-sensitive directions.
-
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.
-
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%.
-
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.
-
Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning
Math reasoning gains in LLMs rarely transfer to general domains; RL tuning generalizes while SFT causes forgetting and representation drift.
-
Entropy-Gradient Inversion: Moving Toward Internal Mechanism of Large Reasoning Models
Defines Entropy-Gradient Inversion as a geometric fingerprint of LRM reasoning and introduces CorR-PO to embed it in RL reward regularization, reporting improved benchmark performance.
-
DyGRO-VLA: Cross-Task Scaling of Vision-Language-Action Models via Dynamic Grouped Residual Optimization
DyGRO-VLA is a two-stage optimization framework for cross-task scaling of Vision-Language-Action models via dynamic grouped residual optimization in RL.
-
Learning to Foresee: Unveiling the Unlocking Efficiency of On-Policy Distillation
On-policy distillation gains efficiency from early foresight in module allocation and update directions, which the proposed EffOPD method exploits for 3x faster training with comparable performance.
-
DACA-GRPO: Denoising-Aware Credit Assignment for Reinforcement Learning in Diffusion Language Models
DACA-GRPO adds denoising-aware credit assignment and bias-reduced likelihood estimation to GRPO, delivering consistent gains up to 36.3pp on math, code, constraint, and schema benchmarks for diffusion LLMs.
- Bad Seeing or Bad Thinking? Rewarding Perception for Multimodal Reasoning
- RAG over Thinking Traces Can Improve Reasoning Tasks