TailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing
cs.CV, cs.AI
Submitted: 2026-11-04
Updated: 2026-11-04
Project page: https://lyricccco.github.io/tailsteer
Terminology
Sources
- MONET: A Massive, Open, Non-redundant and Enriched Text-to-image dataset
- FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space
- SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture
- MT-EditFlow: Reinforcement Learning for Multi-Turn Image Editing with Flow Matching
- BizGenEval: A Systematic Benchmark for Commercial Visual Content Generation
- Step1X-Edit: A Practical Framework for General Image Editing
- Why Do DiT Editors Drift? Plug-and-Play Low Frequency Alignment in VAE Latent Space
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs