# Beyond Attention Masks: Instruction Anchoring for Efficient In-Context Diffusion Generation

arXiv: https://arxiv.org/abs/2608.21229 · HTML: https://mlsanity.com/paper/2608.21229/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models.
- Builds on 32 of its 41 references that are in the map.
- First posted on arXiv 21 Aug 2026.

## Placed in

- Diffusion models / [Diffusion foundations and sampling](https://mlsanity.com/diffusion/diffusion-foundations-and-sampling/)

## Builds on (most cited first, 30 of 32)

- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Scaling Rectified Flow Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2403.03206/) (2024)
- [Consistency Models](https://mlsanity.com/paper/2303.01469/) (2023)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion](https://mlsanity.com/paper/2506.08009/) (2025)
- [Policy Distillation](https://mlsanity.com/paper/1511.06295/) (2015)
- [From Slow Bidirectional to Fast Autoregressive Video Diffusion Models](https://mlsanity.com/paper/2412.07772/) (2024)
- [OminiControl: Minimal and Universal Control for Diffusion Transformer](https://mlsanity.com/paper/2411.15098/) (2024)
- [Seed-TTS: A Family of High-Quality Versatile Speech Generation Models](https://mlsanity.com/paper/2406.02430/) (2024)
- [Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model](https://mlsanity.com/paper/2411.19108/) (2024)
- [Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity](https://mlsanity.com/paper/2502.01776/) (2025)
- [Self-Forcing++: Towards Minute-Scale High-Quality Video Generation](https://mlsanity.com/paper/2510.02283/) (2025)
- [SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models](https://mlsanity.com/paper/2411.05007/) (2024)
- [OmniEdit: Building Image Editing Generalist Models Through Specialist Supervision](https://mlsanity.com/paper/2411.07199/) (2024)
- [Learning-to-Cache: Accelerating Diffusion Transformer via Layer Caching](https://mlsanity.com/paper/2406.01733/) (2024)
- [MiniLLM: On-Policy Distillation of Large Language Models](https://mlsanity.com/paper/2306.08543/) (2023)
- [SpargeAttention: Accurate and Training-free Sparse Attention Accelerating Any Model Inference](https://mlsanity.com/paper/2502.18137/) (2025)
- [EasyControl: Adding Efficient and Flexible Control for Diffusion Transformer](https://mlsanity.com/paper/2503.07027/) (2025)
- [SageAttention2: Efficient Attention with Thorough Outlier Smoothing and Per-thread INT4 Quantization](https://mlsanity.com/paper/2411.10958/) (2024)
- [DistiLLM-2: A Contrastive Approach Boosts the Distillation of LLMs](https://mlsanity.com/paper/2503.07067/) (2025)
- [Radial Attention: O(nlog n) Sparse Attention with Energy Decay for Long Video Generation](https://mlsanity.com/paper/2506.19852/) (2025)
- [OminiControl2: Efficient Conditioning for Diffusion Transformers](https://mlsanity.com/paper/2503.08280/) (2025)
- [OpenVE-3M: A Large-Scale High-Quality Dataset for Instruction-Guided Video Editing](https://mlsanity.com/paper/2512.07826/) (2025)
- [VMoBA: Mixture-of-Block Attention for Video Diffusion Models](https://mlsanity.com/paper/2506.23858/) (2025)
- [CLEAR: Conv-Like Linearization Revs Pre-Trained Diffusion Transformers Up](https://mlsanity.com/paper/2412.16112/) (2024)
- [SmoothCache: A Universal Inference Acceleration Technique for Diffusion Transformers](https://mlsanity.com/paper/2411.10510/) (2024)
- [SpargeAttention2: Trainable Sparse Attention via Hybrid Top-k+Top-p Masking and Distillation Fine-Tuning](https://mlsanity.com/paper/2602.13515/) (2026)
- [BWCache: Accelerating Video Diffusion Transformers through Block-Wise Caching](https://mlsanity.com/paper/2509.13789/) (2025)
- [EDiT: Efficient Diffusion Transformers with Linear Compressed Attention](https://mlsanity.com/paper/2503.16726/) (2025)

## Abstract (arXiv)

In-context diffusion transformers concatenate instruction, target, and reference tokens into a single sequence for joint attention. Reference-side computation must therefore be repeated at every denoising step, with the cost growing rapidly as more references are added. Decoupling reference tokens from the target enables exact key-value reuse across denoising steps, but prevents the references from attending to the instruction, degrading instruction following and reference fidelity. This trade-off cannot be resolved through attention-mask design alone. We introduce AnchorCache, a parameter-free token-layout and attention-mask co-design that inserts static text anchors. These anchors condition the reference representations on the instruction during cache construction, after which the resulting reference keys and values can be reused exactly across denoising steps. To recover the quality initially lost through this structural conversion, we apply teacher-forced velocity distillation followed by a short on-policy stage that queries the teacher at student-visited states. To our knowledge, this is the first use of on-policy distillation for architectural recovery in diffusion models. Across benchmarks spanning image, speech, and video generation, AnchorCache matches full-attention quality. Its efficiency gains increase with the reference-context size, reaching a 6.40x speedup in diffusion transformer inference.