Open Data Synthesis For Deep Research
Paper
• 2509.00375 • Published
• 72
HF Daily
-
-
Beyond Correctness: Harmonizing Process and Outcome Rewards through RL Training
Paper • 2509.03403 • Published • 23 -
LMEnt: A Suite for Analyzing Knowledge in Language Models from Pretraining Data to Representations
Paper • 2509.03405 • Published • 24 -
SATQuest: A Verifier for Logical Reasoning Evaluation and Reinforcement Fine-Tuning of LLMs
Paper • 2509.00930 • Published • 5 -
Drivel-ology: Challenging LLMs with Interpreting Nonsense with Depth
Paper • 2509.03867 • Published • 211 -
Towards a Unified View of Large Language Model Post-Training
Paper • 2509.04419 • Published • 76 -
Inverse IFEval: Can LLMs Unlearn Stubborn Training Conventions to Follow Real Instructions?
Paper • 2509.04292 • Published • 58 -
Delta Activations: A Representation for Finetuned Large Language Models
Paper • 2509.04442 • Published • 7 -
Why Language Models Hallucinate
Paper • 2509.04664 • Published • 196 -
Set Block Decoding is a Language Model Inference Accelerator
Paper • 2509.04185 • Published • 54 -
Bootstrapping Task Spaces for Self-Improvement
Paper • 2509.04575 • Published • 6 -
On Robustness and Reliability of Benchmark-Based Evaluation of LLMs
Paper • 2509.04013 • Published • 4 -
Reverse-Engineered Reasoning for Open-Ended Generation
Paper • 2509.06160 • Published • 149 -
Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models
Paper • 2509.06949 • Published • 56 -
Reinforcement Learning Foundations for Deep Research Systems: A Survey
Paper • 2509.06733 • Published • 32 -
Scaling up Multi-Turn Off-Policy RL and Multi-Agent Tree Search for LLM Step-Provers
Paper • 2509.06493 • Published • 12 -
SFR-DeepResearch: Towards Effective Reinforcement Learning for Autonomously Reasoning Single Agents
Paper • 2509.06283 • Published • 17 -
Test-Time Scaling in Reasoning Models Is Not Effective for Knowledge-Intensive Tasks Yet
Paper • 2509.06861 • Published • 9 -
R^textbf{2AI}: Towards Resistant and Resilient AI in an Evolving World
Paper • 2509.06786 • Published • 3 -
Parallel-R1: Towards Parallel Thinking via Reinforcement Learning
Paper • 2509.07980 • Published • 105 -
Sharing is Caring: Efficient LM Post-Training with Collective RL Experience Sharing
Paper • 2509.08721 • Published • 662 -
Staying in the Sweet Spot: Responsive Reasoning Evolution via Capability-Adaptive Hint Scaffolding
Paper • 2509.06923 • Published • 22 -
Emergent Hierarchical Reasoning in LLMs through Reinforcement Learning
Paper • 2509.03646 • Published • 33 -
ΔL Normalization: Rethink Loss Aggregation in RLVR
Paper • 2509.07558 • Published • 7 -
From Noise to Narrative: Tracing the Origins of Hallucinations in Transformers
Paper • 2509.06938 • Published • 5 -
A Survey of Reinforcement Learning for Large Reasoning Models
Paper • 2509.08827 • Published • 190 -
CDE: Curiosity-Driven Exploration for Efficient Reinforcement Learning in Large Language Models
Paper • 2509.09675 • Published • 28 -
The Majority is not always right: RL training for solution aggregation
Paper • 2509.06870 • Published • 15 -
Statistical Methods in Generative AI
Paper • 2509.07054 • Published • 11 -
MachineLearningLM: Continued Pretraining Language Models on Millions of Synthetic Tabular Prediction Tasks Scales In-Context ML
Paper • 2509.06806 • Published • 64 -
The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMs
Paper • 2509.09677 • Published • 35 -
Virtual Agent Economies
Paper • 2509.10147 • Published • 27 -
Single-stream Policy Optimization
Paper • 2509.13232 • Published • 34 -
EconProver: Towards More Economical Test-Time Scaling for Automated Theorem Proving
Paper • 2509.12603 • Published • 9 -
Towards General Agentic Intelligence via Environment Scaling
Paper • 2509.13311 • Published • 72 -
Scrub It Out! Erasing Sensitive Memorization in Code Language Models via Machine Unlearning
Paper • 2509.13755 • Published • 19 -
THOR: Tool-Integrated Hierarchical Optimization via RL for Mathematical Reasoning
Paper • 2509.13761 • Published • 16 -
FlowRL: Matching Reward Distributions for LLM Reasoning
Paper • 2509.15207 • Published • 116 -
Reasoning over Boundaries: Enhancing Specification Alignment via Test-time Delibration
Paper • 2509.14760 • Published • 53 -
Evolving Language Models without Labels: Majority Drives Selection, Novelty Promotes Variation
Paper • 2509.15194 • Published • 33 -
Latent Zoning Network: A Unified Principle for Generative Modeling, Representation Learning, and Classification
Paper • 2509.15591 • Published • 45 -
LIMI: Less is More for Agency
Paper • 2509.17567 • Published • 104 -
GeoPQA: Bridging the Visual Perception Gap in MLLMs for Geometric Reasoning
Paper • 2509.17437 • Published • 17 -
DiffusionNFT: Online Diffusion Reinforcement with Forward Process
Paper • 2509.16117 • Published • 22 -
Analyzing the Effects of Supervised Fine-Tuning on Model Knowledge from Token and Parameter Levels
Paper • 2509.16596 • Published • 14 -
Reasoning Core: A Scalable RL Environment for LLM Symbolic Reasoning
Paper • 2509.18083 • Published • 5 -
Adaptive Kernel Design for Bayesian Optimization Is a Piece of CAKE with LLMs
Paper • 2509.17998 • Published • 1 -
Reinforcement Learning on Pre-Training Data
Paper • 2509.19249 • Published • 67 -
MAPO: Mixed Advantage Policy Optimization
Paper • 2509.18849 • Published • 27 -
What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoT
Paper • 2509.19284 • Published • 23 -
SIM-CoT: Supervised Implicit Chain-of-Thought
Paper • 2509.20317 • Published • 42 -
EmbeddingGemma: Powerful and Lightweight Text Representations
Paper • 2509.20354 • Published • 48 -
Video models are zero-shot learners and reasoners
Paper • 2509.20328 • Published • 100 -
Mixture of Thoughts: Learning to Aggregate What Experts Think, Not Just What They Say
Paper • 2509.21164 • Published • 9 -
VCRL: Variance-based Curriculum Reinforcement Learning for Large Language Models
Paper • 2509.19803 • Published • 120 -
SciReasoner: Laying the Scientific Reasoning Ground Across Disciplines
Paper • 2509.21320 • Published • 101 -
Tree Search for LLM Agent Reinforcement Learning
Paper • 2509.21240 • Published • 92 -
CE-GPPO: Controlling Entropy via Gradient-Preserving Clipping Policy Optimization in Reinforcement Learning
Paper • 2509.20712 • Published • 19 -
Thinking Augmented Pre-training
Paper • 2509.20186 • Published • 23 -
ScaleDiff: Scaling Difficult Problems for Advanced Mathematical Reasoning
Paper • 2509.21070 • Published • 9 -
EPO: Entropy-regularized Policy Optimization for LLM Agents Reinforcement Learning
Paper • 2509.22576 • Published • 135 -
Quantile Advantage Estimation for Entropy-Safe Reasoning
Paper • 2509.22611 • Published • 118 -
Variational Reasoning for Language Models
Paper • 2509.22637 • Published • 69 -
Language Models Can Learn from Verbal Feedback Without Scalar Rewards
Paper • 2509.22638 • Published • 70 -
No Prompt Left Behind: Exploiting Zero-Variance Prompts in LLM Reinforcement Learning via Entropy-Guided Advantage Shaping
Paper • 2509.21880 • Published • 53 -
PromptCoT 2.0: Scaling Prompt Synthesis for Large Language Model Reasoning
Paper • 2509.19894 • Published • 34 -
HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Models
Paper • 2509.22300 • Published • 4 -
SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention
Paper • 2509.24006 • Published • 118 -
Multiplayer Nash Preference Optimization
Paper • 2509.23102 • Published • 62 -
Beyond the Exploration-Exploitation Trade-off: A Hidden State Approach for LLM Reasoning in RLVR
Paper • 2509.23808 • Published • 47 -
Sequential Diffusion Language Models
Paper • 2509.24007 • Published • 46 -
When Does Reasoning Matter? A Controlled Study of Reasoning's Contribution to Model Performance
Paper • 2509.22193 • Published • 38 -
SparseD: Sparse Attention for Diffusion Language Models
Paper • 2509.24014 • Published • 31 -
Random Policy Valuation is Enough for LLM Reasoning with Verifiable Rewards
Paper • 2509.24981 • Published • 29 -
The Era of Real-World Human Interaction: RL from User Conversations
Paper • 2509.25137 • Published • 19 -
Toward Effective Tool-Integrated Reasoning via Self-Evolved Preference Learning
Paper • 2509.23285 • Published • 14 -
GRPO-MA: Multi-Answer Generation in GRPO for Stable and Efficient Chain-of-Thought Training
Paper • 2509.24494 • Published • 11 -
The Dragon Hatchling: The Missing Link between the Transformer and Models of the Brain
Paper • 2509.26507 • Published • 547 -
TruthRL: Incentivizing Truthful LLMs via Reinforcement Learning
Paper • 2509.25760 • Published • 55 -
Thinking-Free Policy Initialization Makes Distilled Reasoning Models More Effective and Efficient Reasoners
Paper • 2509.26226 • Published • 34 -
Thinking Sparks!: Emergent Attention Heads in Reasoning Models During Post Training
Paper • 2509.25758 • Published • 23 -
Mem-α: Learning Memory Construction via Reinforcement Learning
Paper • 2509.25911 • Published • 15 -
Attention as a Compass: Efficient Exploration for Process-Supervised RL in Reasoning Models
Paper • 2509.26628 • Published • 17 -
InfoAgent: Advancing Autonomous Information-Seeking Agents
Paper • 2509.25189 • Published • 13 -
Benefits and Pitfalls of Reinforcement Learning for Language Model Planning: A Theoretical Perspective
Paper • 2509.22613 • Published • 10 -
Specialization after Generalization: Towards Understanding Test-Time Training in Foundation Models
Paper • 2509.24510 • Published • 5 -
DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search
Paper • 2509.25454 • Published • 146 -
GEM: A Gym for Agentic LLMs
Paper • 2510.01051 • Published • 90 -
Knapsack RL: Unlocking Exploration of LLMs via Optimizing Budget Allocation
Paper • 2509.25849 • Published • 48 -
It Takes Two: Your GRPO Is Secretly DPO
Paper • 2510.00977 • Published • 32 -
ACON: Optimizing Context Compression for Long-horizon LLM Agents
Paper • 2510.00615 • Published • 34 -
BroRL: Scaling Reinforcement Learning via Broadened Exploration
Paper • 2510.01180 • Published • 20 -
Making, not Taking, the Best of N
Paper • 2510.00931 • Published • 10 -
CurES: From Gradient Analysis to Efficient Curriculum Learning for Reasoning LLMs
Paper • 2510.01037 • Published • 2 -
LongCodeZip: Compress Long Context for Code Language Models
Paper • 2510.00446 • Published • 107 -
ExGRPO: Learning to Reason from Experience
Paper • 2510.02245 • Published • 80 -
Interactive Training: Feedback-Driven Neural Network Optimization
Paper • 2510.02297 • Published • 43 -
RLP: Reinforcement as a Pretraining Objective
Paper • 2510.01265 • Published • 44 -
Aristotle: IMO-level Automated Theorem Proving
Paper • 2510.01346 • Published • 17 -
RLAD: Training LLMs to Discover Abstractions for Solving Reasoning Problems
Paper • 2510.02263 • Published • 9 -
Apriel-1.5-15b-Thinker
Paper • 2510.01141 • Published • 121 -
Large Reasoning Models Learn Better Alignment from Flawed Thinking
Paper • 2510.00938 • Published • 59 -
Self-Improvement in Multimodal Large Language Models: A Survey
Paper • 2510.02665 • Published • 21 -
Continuously Augmented Discrete Diffusion model for Categorical Generative Modeling
Paper • 2510.01329 • Published • 6 -
Pretraining with hierarchical memories: separating long-tail and common knowledge
Paper • 2510.02375 • Published • 6 -
A Practitioner's Guide to Multi-turn Agentic Reinforcement Learning
Paper • 2510.01132 • Published • 6 -
Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models
Paper • 2510.04618 • Published • 129 -
Paper2Video: Automatic Video Generation from Scientific Papers
Paper • 2510.05096 • Published • 119 -
MITS: Enhanced Tree Search Reasoning for LLMs via Pointwise Mutual Information
Paper • 2510.03632 • Published • 42 -
Hybrid Architectures for Language Models: Systematic Analysis and Design Insights
Paper • 2510.04800 • Published • 37 -
Front-Loading Reasoning: The Synergy between Pretraining and Post-Training Data
Paper • 2510.03264 • Published • 25 -
Less is More: Recursive Reasoning with Tiny Networks
Paper • 2510.04871 • Published • 509 -
In-the-Flow Agentic System Optimization for Effective Planning and Tool Use
Paper • 2510.05592 • Published • 107 -
MixReasoning: Switching Modes to Think
Paper • 2510.06052 • Published • 22 -
Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning
Paper • 2510.04081 • Published • 23 -
Cache-to-Cache: Direct Semantic Communication Between Large Language Models
Paper • 2510.03215 • Published • 98 -
Lumina-DiMOO: An Omni Diffusion Large Language Model for Multi-Modal Generation and Understanding
Paper • 2510.06308 • Published • 55 -
Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer
Paper • 2510.06590 • Published • 77 -
Multi-Agent Tool-Integrated Policy Optimization
Paper • 2510.04678 • Published • 31 -
Agent Learning via Early Experience
Paper • 2510.08558 • Published • 273 -
Meta-Awareness Enhances Reasoning Models: Self-Alignment Reinforcement Learning
Paper • 2510.03259 • Published • 57 -
When Thoughts Meet Facts: Reusable Reasoning for Long-Context LMs
Paper • 2510.07499 • Published • 48 -
Low-probability Tokens Sustain Exploration in Reinforcement Learning with Verifiable Reward
Paper • 2510.03222 • Published • 75 -
QeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMs
Paper • 2510.11696 • Published • 181 -
Diffusion Transformers with Representation Autoencoders
Paper • 2510.11690 • Published • 166 -
RLFR: Extending Reinforcement Learning for LLMs with Flow Environment
Paper • 2510.10201 • Published • 36 -
Demystifying Reinforcement Learning in Agentic Reasoning
Paper • 2510.11701 • Published • 33 -
Don't Just Fine-tune the Agent, Tune the Environment
Paper • 2510.10197 • Published • 30 -
Memory as Action: Autonomous Context Curation for Long-Horizon Agentic Tasks
Paper • 2510.12635 • Published • 17 -
Attention Illuminates LLM Reasoning: The Preplan-and-Anchor Rhythm Enables Fine-Grained Policy Optimization
Paper • 2510.13554 • Published • 58 -
Stronger Together: On-Policy Reinforcement Learning for Collaborative LLMs
Paper • 2510.11062 • Published • 29 -
Tracing the Traces: Latent Temporal Signals for Efficient and Accurate Reasoning
Paper • 2510.10494 • Published • 2 -
Agentic Entropy-Balanced Policy Optimization
Paper • 2510.14545 • Published • 106 -
LaSeR: Reinforcement Learning with Last-Token Self-Rewarding
Paper • 2510.14943 • Published • 40 -
Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn LLM Agents
Paper • 2510.14967 • Published • 34 -
LLMs Can Get "Brain Rot"!
Paper • 2510.13928 • Published • 23 -
LLM-guided Hierarchical Retrieval
Paper • 2510.13217 • Published • 21 -
Large Language Models Do NOT Really Know What They Don't Know
Paper • 2510.09033 • Published • 17