Downloads · 30 days
0
trm-model-arc-agi-1
alphaXiv/trm-model-arc-agi-1
trm-model-arc-agi-1 is a machine learning model from alphaXiv. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
This is a Tiny Recursive Model (TRM) fine-tuned for solving Abstract Reasoning Challenge (ARC-AGI) tasks. The model performs abstract reasoning to predict output grids from input grids.
Access
Public
Updated Oct 22, 2025
Repo size
3.6 GB
Likes
5
Public
Repo makeup
Click a slice to open those files.
Other3.6 GB · 100%
At a glance
- License
- mit
- Access
- Public
- Created
- Oct 20, 2025
- Updated
- Oct 22, 2025
- SHA
- 35086baf
Try a prompt
Datasets
Model card
From the Hugging Face model README
TRM Model for ARC-AGI-1
Model Description
This is a Tiny Recursive Model (TRM) fine-tuned for solving Abstract Reasoning Challenge (ARC-AGI) tasks. The model performs abstract reasoning to predict output grids from input grids.
- Developed by: alphaXiv
- Model type: TRM-Attention
- Language(s) (NLP): N/A (grid-based reasoning)
- License: MIT
- Finetuned from model: Custom TRM architecture
Intended Use
Primary Use
This model is designed to solve ARC-AGI tasks by predicting the correct output grid transformation based on input grid patterns.
Out-of-Scope Use
Not intended for general NLP tasks, image generation, or other reasoning domains.
Limitations and Bias
- Trained only on ARC-AGI training and evaluation sets
- May not generalize to novel abstract reasoning tasks
- Performance limited by training data diversity
Training Data
The model was trained on the ARC-AGI dataset, which includes:
- Input-output grid pairs
- Various transformation patterns
- Training and evaluation splits
Evaluation Results
| Metric | Claimed | Achieved |
|---|---|---|
| Pass@2 | 44.6% | 43.00% ± 0.16% |
Results from independent reproduction study.
Repository
- License
- mit
- Languages
- en
- Created
- Oct 20, 2025
- Updated
- Oct 22, 2025