# Never Too Late for Force: Accelerating VLA Post-Training with Reactive Force Injection

arXiv: https://arxiv.org/abs/2607.14236 · HTML: https://mlsanity.com/paper/2607.14236/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Tactile, force-aware and humanoid VLA, one of 13 directions in Vision-language-action models.
- Builds on 15 of its 32 references that are in the map.
- First posted on arXiv 15 Jul 2026.

## Placed in

- Vision-language-action models / [Tactile, force-aware and humanoid VLA](https://mlsanity.com/vla/tactile-force-aware-and-humanoid-vla/)

## Builds on (most cited first, 15 of 15)

- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation](https://mlsanity.com/paper/2503.02881/) (2025)
- [Multi-Task Interactive Robot Fleet Learning with Visual World Models](https://mlsanity.com/paper/2410.22689/) (2024)
- [FACTR: Force-Attending Curriculum Training for Contact-Rich Policy Learning](https://mlsanity.com/paper/2502.17432/) (2025)
- [TA-VLA: Elucidating the Design Space of Torque-aware Vision-Language-Action Models](https://mlsanity.com/paper/2509.07962/) (2025)
- [ForceVLA2: Unleashing Hybrid Force-Position Control with Force Awareness for Contact-Rich Manipulation](https://mlsanity.com/paper/2603.15169/) (2026)
- [SOP: A Scalable Online Post-Training System for Vision-Language-Action Models](https://mlsanity.com/paper/2601.03044/) (2026)
- [FAVLA: A Force-Adaptive Fast-Slow VLA model for Contact-Rich Robotic Manipulation](https://mlsanity.com/paper/2602.23648/) (2026)
- [Force Policy: Learning Hybrid Force-Position Control Policy under Interaction Frame for Contact-Rich Manipulation](https://mlsanity.com/paper/2602.22088/) (2026)
- [ImplicitRDP: An End-to-End Visual-Force Diffusion Policy with Structural Slow-Fast Learning](https://mlsanity.com/paper/2512.10946/) (2025)
- [SOE: Sample-Efficient Robot Policy Self-Improvement via On-Manifold Exploration](https://mlsanity.com/paper/2509.19292/) (2025)
- [RoboPocket: Improve Robot Policies Instantly with Your Phone](https://mlsanity.com/paper/2603.05504/) (2026)

## Built on by (most cited first, 3 of 3)

- [GRAFT: Grounded and Efficient Online Reinforcement Adaptation for Fine-Grained Robot Manipulation](https://mlsanity.com/paper/2608.27079/) (2026)
- [ForceDelta-VLA: Distilling Force-Conditioned ActionCorrections for Contact-Rich Manipulation](https://mlsanity.com/paper/2609.18242/) (2026)
- [VT-Bridge: Bridging Pretrained Foundation VLAs to VTLAs via Lightweight Residual Adaptation](https://mlsanity.com/paper/2609.22606/) (2026)

## Abstract (arXiv)

Pretrained vision-language-action (VLA) policies provide strong language-conditioned manipulation knowledge, but they remain largely vision-driven and can struggle once manipulation enters contact states where the scene is occluded, depth is ambiguous, or small force errors push execution off the offline demonstration distribution. We present LIFT (Late Reactive Injection of Force for VLA Post-Training), a force-aware post-training framework that adds contact reactivity to a pretrained VLA policy while preserving its output at initialization. LIFT grafts a reactive action expert beside the original action expert, initializes it from pretrained action weights, and injects recent 6D end-effector force through causal force memory and zero-initialized cross attention, enabling actions to be refreshed during execution. To address the policy-dependent distribution shift of contact feedback, LIFT further couples reactive force injection with an online DAgger loop that trains on a mixture of offline task-alignment data and human-corrected online rollouts. Across towel folding, book insertion, and Hanoi ring placement, LIFT learns faster and reaches higher peak and final performance than vision-only DAgger, while ablations show that reactive force memory and online corrective data are both important for robust contact-rich manipulation. Our code is publicly available at https://github.com/y-wng/lift