Steve Helix: Created page with "= Tiny Recursive Model (TRM) = Tiny Recursive Model diagram (from Samsung SAIL Montréal) '''Tiny Recursive Model (TRM)''' is a compact neural architecture proposed by Alexia Jolicoeur-Martineau at Samsung SAIL Montréal that demonstrates advanced reasoning capabilities despite being ~10,000x smaller than most large language models. == Overview == TRM is designed around the principle of **recursi..."

2025-10-08T18:13:44Z

Created page with "= Tiny Recursive Model (TRM) = thumb|right|400px|Tiny Recursive Model diagram (from Samsung SAIL Montréal) '''Tiny Recursive Model (TRM)''' is a compact neural architecture proposed by Alexia Jolicoeur-Martineau at Samsung SAIL Montréal that demonstrates advanced reasoning capabilities despite being ~10,000x smaller than most large language models. == Overview == TRM is designed around the principle of **recursi..."

New page

= Tiny Recursive Model (TRM) =

[[File:TRM_Architecture.png|thumb|right|400px|Tiny Recursive Model diagram (from Samsung SAIL Montréal)]]

'''Tiny Recursive Model (TRM)''' is a compact neural architecture proposed by [[User:AlexiaJolicoeur|Alexia Jolicoeur-Martineau]] at Samsung SAIL Montréal that demonstrates advanced reasoning capabilities despite being ~10,000x smaller than most large language models.

== Overview ==
TRM is designed around the principle of **recursive self-refinement**. Unlike typical LLMs that predict responses token-by-token in a single pass, TRM generates an initial draft, constructs an internal “latent reasoning” state (scratchpad), and then improves its own answer over up to 16 iterations. This mimics meta-cognition in a compact form.

== Key Features ==
* **Tiny Model Size**: ~7M parameters
* **Deep Reasoning Loop**: Recursively improves answers across time steps
* **Superior Test Accuracy**:
* 45% on ARC-AGI-1
* 8% on ARC-AGI-2
* Outperforms Gemini 2.5 Pro, DeepSeek R1, and O3-mini on these benchmarks
* **Minimal Training Data**: Trained on only ~1000 examples

== Architecture Summary ==
* Two small networks: a **Reasoning MLP** and a **Prediction MLP**
* Reasoning loop feeds latent variables into the Prediction MLP
* Uses self-attention + scratchpad updates per reasoning step
* Final output selected after convergence or max iteration (n=16)

== Comparison to LLMs ==
{| class="wikitable"
! Feature !! TRM !! Large Language Models
|-
| Parameters || ~7M || 70B–175B+
| Reasoning Style || Iterative, recursive || Single-pass prediction
| ARC-AGI Performance || ↑ Better on logic tasks || ↓ Weaker generalization
| Memory Footprint || Extremely low || High
| Explainability || Easier to trace steps || Often opaque
| Use Case Fit || Logic puzzles, verifiable computation || Open-ended language, wide context
|}

== Relevance to Helix ==
TRM offers a feasible lightweight module for tasks that require:
* Deterministic verification
* Local reasoning on edge devices
* Constrained-resource environments
* Transparent reasoning chains (auditable steps)

Possible integration points:
* [[Helix_Models_Overview]]
* [[Helix_Core_Ethos|Helix Core Ethos]] (as a trusted verification subagent)
* [[QSR_Runbook]] for recursive policy validation
* [[Runbooks_&_Protocols]]: as a fallback verifier model

== Resources ==
* [https://github.com/SamsungSAILMontreal/TinyRecursiveModels GitHub – Tiny Recursive Models]
* [https://arxiv.org/abs/2510.04871v1 ArXiv Paper (2025-10-06)]
* Related: [[Helix LLaMA:8B – Operating Ontology]], [[Khronos: Experimental Runtime Agent]]

== License ==
MIT – freely reusable with attribution

== Status ==
{{Status|experimental}}

[[Category:Helix Models]]
[[Category:AI Research]]
[[Category:Recursive Reasoning]]

Tiny Recursive Model (TRM) - Revision history