REVIEW 13 cited by
MM-RLHF: The Next Step Forward in Multimodal LLM Alignment
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
MM-RLHF: The Next Step Forward in Multimodal LLM Alignment
read the original abstract
Despite notable advancements in Multimodal Large Language Models (MLLMs), most state-of-the-art models have not undergone thorough alignment with human preferences. This gap exists because current alignment research has primarily achieved progress in specific areas (e.g., hallucination reduction), while the broader question of whether aligning models with human preferences can systematically enhance MLLM capability remains largely unexplored. To this end, we introduce MM-RLHF, a dataset containing $\mathbf{120k}$ fine-grained, human-annotated preference comparison pairs. This dataset represents a substantial advancement over existing resources, offering superior size, diversity, annotation granularity, and quality. Leveraging this dataset, we propose several key innovations to improve both the quality of reward models and the efficiency of alignment algorithms. Notably, we introduce a Critique-Based Reward Model, which generates critiques of model outputs before assigning scores, offering enhanced interpretability and more informative feedback compared to traditional scalar reward mechanisms. Additionally, we propose Dynamic Reward Scaling, a method that adjusts the loss weight of each sample according to the reward signal, thereby optimizing the use of high-quality comparison pairs. Our approach is rigorously evaluated across $\mathbf{10}$ distinct dimensions and $\mathbf{27}$ benchmarks, with results demonstrating significant and consistent improvements in model performance. Specifically, fine-tuning LLaVA-ov-7B with MM-RLHF and our alignment algorithm leads to a $\mathbf{19.5}$% increase in conversational abilities and a $\mathbf{60}$% improvement in safety. We have open-sourced the preference dataset, reward model, training and evaluation code, as well as reward modeling and safety benchmarks. For more details, please visit our project page: https://mm-rlhf.github.io.
Forward citations
Cited by 13 Pith papers
-
LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning
LatentOmni proposes a latent-space cross-modal reasoning framework that uses feature-level supervision and Omni-Sync Position Embedding to align and synchronize audio-visual latents, supported by a new 35K interleaved...
-
Culture-Aware Humorous Captioning: Multimodal Humor Generation across Cultural Contexts
Introduces culture-aware humorous captioning task and staged alignment framework that improves contextual fit and balances image relevance with humor in multimodal LLMs.
-
Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation
DeP mitigates MLLM hallucinations by dynamically perturbing text prompts to identify and reinforce stable visual evidence regions while counteracting language prior biases using attention variance and logit statistics.
-
Multimodal Reward Hacking in Reinforcement Learning
Imperfect multimodal RL rewards systematically create new failures (NRFR > RHR); scaling and answer-aware rewards help but do not eliminate hacking, and unreliable visual verifiers actively increase it.
-
Artifact-Bench: Evaluating MLLMs on Detecting and Assessing the Artifacts of AI-Generated Videos
Artifact-Bench supplies a three-level artifact taxonomy and three evaluation tasks that show 19 MLLMs perform near or below random on AI-video realism detection and reasoning.
-
DeltaRubric: Generative Multimodal Reward Modeling via Joint Planning and Verification
DeltaRubric decomposes multimodal preference evaluation into self-generated planning and verification steps within a single model, producing large accuracy improvements on VL-RewardBench via multi-role reinforcement learning.
-
Building a Precise Video Language with Human-AI Oversight
CHAI framework pairs AI pre-captions with expert human critiques to produce precise video descriptions, enabling open models to outperform closed ones like Gemini-3.1-Pro and improve fine-grained control in video gene...
-
DriveReward: A Comprehensive Dataset and Generative Vision-Language Reward Model for Autonomous Driving
Creates DriveReward dataset with counterfactual annotations and a 1B VLM reward model that outperforms larger VLMs on driving tasks and matches rule-based rewards in RL and trajectory scoring.
-
Constitutional On-Policy Safe Distillation
COPSD uses a Cross-SFT cold-start followed by constitution-conditioned distillation to achieve stronger safety-helpfulness balance and lower safety tax on reasoning than prior on-policy self-distillation methods.
-
NoisyGRPO: Incentivizing Multimodal CoT Reasoning via Noise Injection and Bayesian Estimation
NoisyGRPO is an RL framework that perturbs visual inputs with Gaussian noise for exploration and computes trajectory advantages via Bayesian posterior fusion of noise prior and reward likelihood to improve multimodal ...
-
Kwai Keye-VL-2.0 Technical Report
Kwai Keye-VL-2.0-30B-A3B is a 30B MoE model with 3B active parameters using DSA adaptation and MOPD distillation that reports SOTA results on video understanding and agent benchmarks.
-
Toward Native Multimodal Modeling: A Roadmap
A roadmap that defines architectural nativity for multimodal models and categorizes them into Multi-to-Text, Multi-to-Target, and Multi-to-Multi types while outlining an industrial pipeline toward unified transformer-...
-
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.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.