Why test-time training? – Rabbitholes
I was first introduced to test-time training (TTT) by Ekin Akyürek et al.’s SoTA performance on ARC-AGI (Akyürek et al. 2024). This idea was nonintuitive but compelling nonetheless, so I structured my deep learning final project around adapting pre-trained language models with Titan modules for long-context memory efficiency. Within an hour of training on a single GPU, updates made at test time had already recovered much of the performance gap between full-fidelity attention and a less expressive version. Almost a year later, I have gone through multiple iterations of this project, trying out new architecture changes and training methods. Still, I find myself asking fundamental questions about this research direction. Why should gradients be computed at test time? What limits on learning should the phrasing of test time imply? Why was I working on extending context lengths when agent harnesses had \compact as a way to summarize the entire token window? Why spend time on TTT when KV cac
I was first introduced to test-time training (TTT) by Ekin Akyürek et al.’s SoTA performance on ARC-AGI (Akyürek et al. 2024). This idea was nonintuitive but compelling nonetheless, so I structured my deep learning final project around adapting pre-trained language models with Titan modules for long-context memory efficiency. Within an hour of training on a single GPU, updates made at test time had already recovered much of the performance gap between full-fidelity attention and a less expressive version. Almost a year later, I have gone through multiple iterations of this project, trying…
saved by
related reading
- [2512.23675] End-to-End Test-Time Training for Long Contextarxiv.org
- Test-Time Training with KV Binding Is Secretly Linear Attentionresearch.nvidia.com
- Reimagining LLM Memory: Using Context as Training Data Unlocks Models That Learn at Test-Time | NVIDIA Technical Blogdeveloper.nvidia.com
- Test-Time Training Done Rightarxiv.org
- Learning to (Learn at Test Time): RNNs with Expressive Hidden Statesarxiv.org
- Noam Brown on X: "Implications of Large-Scale Test-Time Compute" / Xx.com
- Optimizing LLM Test-Time Compute Involves Solving a Meta-RL Problem – Machine Learning Blog | ML@CMU | Carnegie Mellon Universityblog.ml.cmu.edu
- Composer2.pdfcursor.com
- Learning to (Learn at Test Time): RNNs with Expressive Hidden Statesarxiv.org
- [2501.00663] Titans: Learning to Memorize at Test Timearxiv.org
- [2407.04620] Learning to (Learn at Test Time): RNNs with Expressive Hidden Statesar5iv.labs.arxiv.org
- Andrej Karpathy — AGI is still a decade awaydwarkesh.com