Recurrent Looped Transformer: new architecture adds unbounded temporal depth for RL-friendly compute
burkov · x · 2026-09-14
This post introduces the Recurrent Looped Transformer (RLT), a new architecture targeting three weaknesses of standard transformers: state resets at prompt-response boundaries that cause policy-evaluation mismatches in RL, underutilized hardware due to sequential dependencies, and limited latent reasoning depth.
Key design choices:
- Recurrent state passing enables unbounded temporal depth
- Encoder/decoder role separation unlocks hardware optimizations
- Identical state transitions across pretraining, finetuning, sampling, and replay to support reliable RL scaling
More from Research
- NVIDIA Nemotron 3.5 scales LLM self-verification, +9-10 points on agentic benchmarks — sarahcat21 · 2026-10-31
- Unofficial OpenTargets MCP Server Brings Gene-Drug-Disease Data to LLMs — modelcontextprotocol · 2026-10-31
- Researchers find 5 personality structures in LLM weight space, Big Five-like — luke_drago_ · 2026-10-31
- NVS benchmarks are misaligned: proper pose alignment puts 3D Gaussians back at SOTA — kwangmoo_yi · 2026-10-31
- arXiv moderator: submission surge causes months-long holds, new moderators wanted — tdietterich · 2026-10-31
- arXiv policy checks reject class projects, proposals, surveys and position papers — tdietterich · 2026-10-31