AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

Actions as Language: Fine-Tuning VLMs into VLAs Without Catastrophic Forgetting

Asher James Hancock, Xindi Wu, Lihan Zha, Olga Russakovsky, Anirudha Majumdar

ICLR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
represent actions as natural language and fine-tune with LoRA onlyVLM2VLA (Gemma-3-12B-IT) · 12 VQA benchmarksVQA score retention: MME 1182.3 -> 1391.7; MMMU 46.0 -> 42.7Improved
fine-tune VLM into VLA with tokenized actions and full fine-tuning (OpenVLA/ECoT)Prismatic 7B · VQA benchmarksVQA score: MMStar 38.8 -> 0 (OpenVLA)Hurt
replace language actions with least-likely action tokens (VLM2VLA-AT) (ablation)VLM2VLA · VQA benchmarksVQA score: ChartQA 58.3 -> 41.8; MMMU 42.7 -> 45.9No clear effect
represent actions as language instead of least-likely tokens (ablation)VLM2VLA · Real robot OOD, Pick Up the Item Above Ash Ketchumsuccess rate (%): 30% -> 60%Improved
represent actions as language instead of least-likely tokens (ablation)VLM2VLA · Real robot, simple ID taskssuccess rateNo clear effect
represent actions as language instead of least-likely tokens (ablation)VLM2VLA · Real robot multilingual and compositional taskssuccess rateImproved
co-train with robot and web data (pi0.5)pi0.5 · VQA benchmarksVQA score: RealWorldQA 2.7Hurt
use OpenVLA trained on Open-X-Embodiment versus VLM2VLAOpenVLA · Real robot ID pick-up and pick-and-placesuccess rateImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy