Shanghai AI Lab's OpenESM shows energy-steered models can scale as foundation learners
Shanghai-AI-Laboratory · hf · 2026-10-28
Shanghai AI Laboratory introduces Energy-Steered Models (ESMs), integrating energy-guided latent optimization with token generation in a single architecture: the model first optimizes a latent state over the token landscape, then uses it to refine generation, so favorable tokens correspond to lower energy values. Pre-trained on three large-scale corpora and evaluated across 26 benchmarks, ESMs scale effectively and improve foundational language modeling, challenging the view that energy-based models suit only domain-specific tasks. Code and models are open-sourced.
More from Research
- RAX lets a real robot explore for 2 hours and turn learned skills into reusable code — no demos needed — du_yilun · 2026-10-28
- Three approaches to using one model to help build another that makes decisions — TheTuringPost · 2026-10-28
- ML models predict lifespan-extending compounds with up to 80.36% accuracy — Dr_Singularity · 2026-10-28
- MQAScore: fine-grained text-music alignment eval across 5 dimensions via LALM — affige_yang · 2026-10-28
- OpenTac tactile sensors demoed at CoRL 2026 among 17 robot demos — _krishna_murthy · 2026-10-28
- CellTrack-Bench tests VLMs on microscopic cell tracking, exposing reasoning gaps — biohub · 2026-10-28