test-time-training

This paper introduces a trained fast-weight memory mechanism for a 3M-parameter transformer that installs never-trained rules at inference using forward-only test-time training, showing no transfer. The work is fully reproducible on a single RTX 3090.

This paper introduces a behavioral evaluation framework for calibrating claims about deployment-time memory in LLM test-time training, proposing an evidence ladder and explicit baselines to bridge proxy metrics and behavioral evidence.

This paper introduces U-TTT, a U-shaped deep learning model with test-time training layers and dual-domain adaptation for robust PET image denoising under distribution shifts, achieving state-of-the-art performance across different dose levels and scanner types.

Proposes EASE-TTT, a test-time training framework that aligns adaptation with retrieved evidence to improve long-context QA performance in smaller language models.

Alpha-RTL (TTT-RTL) introduces a test-time training framework for RTL hardware optimization, using reinforcement learning with EDA feedback to refine LLM-generated designs. It achieves significant PPA reductions on benchmarks.

This paper proposes a method to convert pretrained Softmax attention models into linear-complexity Test-Time Training (TTT) architectures, achieving comparable text-to-image quality to fine-tuned Softmax models while significantly accelerating inference. The approach is validated by linearizing Stable Diffusion 3.5, resulting in SD3.5-T^5 with 1.32x speedup at 1K resolution.

This paper identifies three threat models for test-time training (TTT) that adversaries can exploit to bypass safety filters in LLMs, achieving high attack success rates. The findings reveal that TTT introduces new vulnerabilities that undermine existing safety guardrails.

TEMPO introduces a test-time training framework that alternates policy refinement with critic recalibration to prevent diversity collapse and sustain performance gains in large reasoning models, boosting AIME 2024 scores for Qwen3-14B from 42.3% to 65.8%.