Enregistré dans:
Détails bibliographiques
Auteurs principaux: Talley, Caleb, Tibrewal, Vedant, Adekunle, Seun, Dong, Weiwen, Wu, Xinyu, Sheikh, Fariha
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2605.01154
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866918477895303168
author Talley, Caleb
Tibrewal, Vedant
Adekunle, Seun
Dong, Weiwen
Wu, Xinyu
Sheikh, Fariha
author_facet Talley, Caleb
Tibrewal, Vedant
Adekunle, Seun
Dong, Weiwen
Wu, Xinyu
Sheikh, Fariha
contents ARC-AGI-2 is a benchmark of human-intuitive visual puzzles that measures a machine's ability to generalize from limited examples, interpret symbolic meaning, and flexibly apply rules in varying contexts. In this paper, we discuss our approach to solving the ARC-AGI-2 puzzles with TinyLM, with additional fine-tuning at test time, including Test-Time-Training (TTT) and Products of Experts (POE). Our model achieves 96.1% accuracy on the training set and 21.7% accuracy on the evaluation set.
format Preprint
id arxiv_https___arxiv_org_abs_2605_01154
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Multi-Perspective Transformers in ARC-AGI-2 Challenge
Talley, Caleb
Tibrewal, Vedant
Adekunle, Seun
Dong, Weiwen
Wu, Xinyu
Sheikh, Fariha
Machine Learning
Artificial Intelligence
ARC-AGI-2 is a benchmark of human-intuitive visual puzzles that measures a machine's ability to generalize from limited examples, interpret symbolic meaning, and flexibly apply rules in varying contexts. In this paper, we discuss our approach to solving the ARC-AGI-2 puzzles with TinyLM, with additional fine-tuning at test time, including Test-Time-Training (TTT) and Products of Experts (POE). Our model achieves 96.1% accuracy on the training set and 21.7% accuracy on the evaluation set.
title Multi-Perspective Transformers in ARC-AGI-2 Challenge
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2605.01154