REVIEW 40 cited by
VisualPRM: An Effective Process Reward Model for Multimodal Reasoning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
We introduce VisualPRM, an advanced multimodal Process Reward Model (PRM) with 8B parameters, which improves the reasoning abilities of existing Multimodal Large Language Models (MLLMs) across different model scales and families with Best-of-N (BoN) evaluation strategies. Specifically, our model improves the reasoning performance of three types of MLLMs and four different model scales. Even when applied to the highly capable InternVL2.5-78B, it achieves a 5.9-point improvement across seven multimodal reasoning benchmarks. Experimental results show that our model exhibits superior performance compared to Outcome Reward Models and Self-Consistency during BoN evaluation. To facilitate the training of multimodal PRMs, we construct a multimodal process supervision dataset VisualPRM400K using an automated data pipeline. For the evaluation of multimodal PRMs, we propose VisualProcessBench, a benchmark with human-annotated step-wise correctness labels, to measure the abilities of PRMs to detect erroneous steps in multimodal reasoning tasks. We hope that our work can inspire more future research and contribute to the development of MLLMs. Our model, data, and benchmark are released in https://internvl.github.io/blog/2025-03-13-VisualPRM/.
Forward citations
Cited by 40 Pith papers
-
ATLAS: Agentic Test-time Learning-to-Allocate Scaling
ATLAS introduces an LLM-orchestrated agentic framework for dynamic test-time scaling via extensible 'explore' actions, achieving higher accuracy with fewer API calls than fixed-workflow baselines on four benchmarks.
-
StemBind: When MLLMs Get Lost Between Rules and Instances in Abstract Visual Reasoning
StemBind benchmark diagnoses MLLM failures in abstract visual reasoning by separating perception, rule induction, and answer selection on shared stems, finding a persistent rule-to-instance binding gap even when perce...
-
PRO-CUA: Process-Reward Optimization for Computer Use Agents
PRO-CUA trains CUAs via decoupled on-policy rollouts and PRM-guided step-level optimization to enable dense credit assignment without expert trajectories or golden answers.
-
T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts
T2I-FactualBench is a new three-tier benchmark for factuality of knowledge-intensive concepts in T2I models, using multi-round VQA evaluation to show SOTA models need improvement.
-
Correcting What You Cannot See: Credit Assignment for Perception Distillation in Multimodal Reasoners
PCD weights perception-span distillation by the product of downstream failure (1−PSR) and teacher–student disagreement, improving average multimodal reasoning benchmark scores by 2.8–4.3 points over uniform on-policy ...
-
Correcting What You Cannot See: Credit Assignment for Perception Distillation in Multimodal Reasoners
Perception-Correction Distillation up-weights teacher supervision on perception spans only when low downstream success and high teacher–student disagreement co-occur, lifting 8B→2B and 32B→8B multimodal averages over ...
-
Reasoning as Intersection: Consensus-Frame Alignment for Visual Focus in Video-MLLMs
CF-GRPO creates a consensus frame prior from intrinsic video cues and aligns it with model frame-use scores via a reward signal to enable evidence-aware reasoning in Video-MLLMs without temporal annotations.
-
SCI-PRM: A Tool Aware Process Reward Model for Scientific Reasoning Verification
Sci-PRM is a tool-aware process reward model trained on the SCIPRM70K dataset to provide fine-grained supervision for scientific reasoning and shown to boost foundation models via Best-of-N selection and RL.
-
PAIR: Prefix-Aware Internal Reward Model for Multi-Turn Agent Optimization
A two-stage probe (hidden-state estimate plus attention-based correction) yields per-step GRPO rewards that survive prefix contamination and beat external-judge and tree-search rewards in the reported benchmarks.
-
PAIR: Prefix-Aware Internal Reward Model for Multi-Turn Agent Optimization
PAIR combines a hidden-state probe with an attention correction to deliver robust step-level rewards for GRPO-based optimization of multi-turn LLM agents, achieving high AUROC on contaminated trajectories at low cost.
-
From Failure to Feedback: Group Revision Unlocks Hard Cases in Object-Level Grounding
A group-revision paradigm for GRPO-based RL fine-tuning of VLMs converts failure responses into improvement signals that refine rewards and advantages, yielding gains on referring segmentation, REC, and counting benchmarks.
-
LLMs Know When They Know, but Do Not Act on It: A Metacognitive Harness for Test-time Scaling
A metacognitive harness uses LLMs' pre- and post-solution self-monitoring signals to control test-time reasoning, raising pooled accuracy from 48.3% to 56.9% on text, code, and multimodal benchmarks.
-
PDCR: Perception-Decomposed Confidence Reward for Vision-Language Reasoning
PDCR improves vision-language reasoning by computing separate normalized confidence advantages for perception steps and reasoning steps after unsupervised decomposition.
-
Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model
SCOLAR addresses information gain collapse in latent visual reasoning by generating independent auxiliary visual tokens from LLM hidden states, extending acceptable CoT length over 30x and achieving +14.12% gains on b...
-
Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model
SCOLAR fixes information gain collapse in latent visual reasoning by generating independent auxiliary visual tokens via a detransformer, extending acceptable CoT length over 30x and delivering +14.12% gains on reasoni...
-
Verification Mirage: Mapping the Reliability Boundary of Self-Verification in Medical VQA
Self-verification in medical VQA creates a verification mirage where verifiers exhibit high error and agreement bias on wrong answers, with reliability strongly conditioned on task type.
-
Chart-FR1: Visual Focus-Driven Fine-Grained Reasoning on Dense Charts
Chart-FR1 uses Focus-CoT for linking reasoning to visual cues and Focus-GRPO reinforcement learning with efficiency rewards to outperform prior MLLMs on dense chart reasoning tasks.
-
V-tableR1: Process-Supervised Multimodal Table Reasoning with Critic-Guided Policy Optimization
V-tableR1 uses a critic VLM for dense step-level feedback and a new PGPO algorithm to shift multimodal table reasoning from pattern matching to verifiable logical steps, achieving SOTA accuracy with a 4B open-source model.
-
Process Reward Models Meet Planning: Generating Precise and Scalable Datasets for Step-Level Rewards
PDDL planning problems are used to generate about one million precise reasoning steps for training Process Reward Models, and adding this data to existing datasets improves LLM performance on both mathematical and non...
-
CoVR-R:Reason-Aware Composed Video Retrieval
Zero-shot LMM reasoning over edit after-effects (states, phases, camera, tempo) plus a new CoVR-R benchmark yields large recall gains on implicit-effect composed video retrieval without task-specific training.
-
EvoLMM: Self-Evolving Large Multimodal Models with Continuous Rewards
A self-evolving multimodal model using continuous self-consistency rewards improves math reasoning by about 2–3% using only raw images, without labels or external reward models.
-
LLaVA-Critic-R1: Your Critic Model is Secretly a Strong Policy Model
RL training on preference-labeled critic data transforms a 7B vision-language model into both a stronger critic and a stronger generative policy, improving average benchmark accuracy by 5.7% and enabling self-critique...
-
InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
InternVL3.5 advances open-source multimodal models with Cascade RL for +16% reasoning gains and ViR for 4x inference speedup, with the 241B model reaching SOTA among open-source MLLMs on multimodal, reasoning, and age...
-
GM-PRM: A Generative Multimodal Process Reward Model for Multimodal Mathematical Reasoning
A generative multimodal process reward model that produces step-level critiques and corrections improves average math accuracy for six multimodal LLMs by 2.9 to 5.9 points under a refinement-based Best-of-N strategy.
-
VL-Cogito: Progressive Curriculum Reinforcement Learning for Advanced Multimodal Reasoning
VL-Cogito, trained with progressive curriculum RL, online difficulty weighting, and dynamic length rewards, matches or beats prior reasoning MLLMs on ten multimodal benchmarks.
-
RewardBench 2: Advancing Reward Model Evaluation
RewardBench 2 is a new benchmark that supplies challenging fresh human prompts for reward model evaluation, yielding lower average scores but higher correlation with downstream best-of-N sampling and RLHF training per...
-
InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models
InternVL3-78B sets a new open-source SOTA of 72.2 on MMMU via native joint multimodal pre-training, V2PE, MPO, and test-time scaling while remaining competitive with proprietary models.
-
OpenVLThinker: Complex Vision-Language Reasoning via Iterative SFT-RL Cycles
Iterative SFT-RL cycles enable a 7B LVLM to develop sophisticated visual chain-of-thought reasoning and improve performance on math and general reasoning benchmarks.
-
Test-Time Scaling for Small VLMs on Multilingual Visual MCQ
On EXAMS-V, token budget and parseability dominate chain count, PRM-guided search, and selectors for small VLMs; the policy model itself yields the largest gain.
-
Test-Time Scaling in Multimodal Foundation Models: A Comprehensive Survey of Generation and Reasoning
A survey of test-time scaling for multimodal foundation models that introduces a three-way taxonomy of sampling, feedback, and search approaches along with applications and benchmarks.
-
CaptchaMind: Training CAPTCHA Solvers via Reinforcement Learning with Explicit Reasoning Supervision
Presents CaptchaBench benchmark and CaptchaMind RL solver achieving 82.9% success on benchmark tasks and 71% on real-world CAPTCHAs via explicit reasoning process supervision.
-
DT2IT-MRM: Debiased Preference Construction and Iterative Training for Multimodal Reward Modeling
DT2IT-MRM proposes a debiased preference construction pipeline, T2I data reformulation, and iterative training to curate multimodal preference data, achieving SOTA on VL-RewardBench, Multimodal RewardBench, and MM-RLH...
-
PaLMR: Towards Faithful Visual Reasoning via Multimodal Process Alignment
By gating the reinforcement-learning reward on a binary visual-fidelity score, PaLMR lifts Qwen2.5-VL-7B to 70.9 on HallusionBench and 59.3 on MMMU, surpassing its GRPO baseline.
-
Physical Plausibility Reasoning via HCM-GRPO: Empowering Compact Model for Superior Performance
A 2B multimodal model trained with HCM-GRPO, a GRPO variant with partial-credit rewards and hard-case oversampling, outperforms larger models on the authors' private physical-plausibility test set.
-
Reinforced Visual Perception with Tools
ReVPT uses GRPO reinforcement learning with a cold-start SFT phase to make Qwen2.5-VL models call visual tools, improving perception benchmarks over SFT and text-only RL baselines.
-
VRPRM: Process Reward Modeling via Visual Reasoning
VRPRM combines visual reasoning with a two-stage SFT-plus-RL strategy to deliver higher-quality process reward modeling using far less annotated data than prior non-thinking PRMs.
-
VRPRM: Process Reward Modeling via Visual Reasoning
VRPRM combines 3.6K CoT-PRM SFT data with 50K non-CoT PRM RL data to train a visual PRM that beats a 400K-data non-thinking PRM and boosts best-of-N accuracy.
-
Improving Multimodal Reasoning via Worst Dimension Optimization
Proposes Worst Dimension Optimization to address equal-weighting limitations in Process Reward Models for multimodal reasoning.
-
From System 1 to System 2: A Survey of Reasoning Large Language Models
The survey organizes the shift of LLMs toward deliberate System 2 reasoning, covering model construction techniques, performance on math and coding benchmarks, and future research directions.
-
Reinforcement Learning from Human Feedback
The book introduces the origins, mathematical setup, and optimization stages of RLHF including reward modeling, reinforcement learning, rejection sampling, and direct alignment algorithms.
Discussion (0). Sign in to comment.