VISA: Agentic Self-Evolving Data Synthesis for Multimodal Instruction Following
cs.CL
Submitted: 2026-08-26
Updated: 2026-08-27
Terminology
Sources
- Qwen3-VL Technical Report
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Empowering Reliable Visual-Centric Instruction Following in MLLMs
- LLaVA-OneVision: Easy Visual Task Transfer
- Reinforcement Learning with Verifiable yet Noisy Rewards under Imperfect Verifiers
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- HybridFlow: A Flexible and Efficient RLHF Framework
- Qwen3.5-Omni Technical Report
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- Reinforcement Learning with Verifiable Rewards Implicitly Incentivizes Correct Reasoning in Base LLMs
- WizardLM: Empowering large pre-trained language models to follow complex instructions
- ReAct: Synergizing Reasoning and Acting in Language Models
- MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
- MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities
- Instruction-Following Evaluation for Large Language Models
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering