ICLR 2026PastOther
Third Workshop on Test-Time Updates (Main Track)
TTU at ICLR 2026 (Main)
- Submission deadline
- Feb 7, 2026, 11:59 UTCOpenReview-synced 2026-02-07 11:59 UTC (as of 2026-06-23) — extensions on OpenReview are applied automatically; verify on the website.
- Submission portal
- OpenReview
- Notes
- Topics were auto-suggested and may be imprecise — edits welcome.
Accepted papers (48)
Fetched from OpenReview (v2) on 2026-06-10.
Adapting in the Dark: Efficient and Stable Test-Time Adaptation for Black-Box Models
An Optimal Transport View of Activation Steering In Masked Diffusion Models
Caravan: Asynchronous Test-Time Adaptation for Faster Inference
Challenges in Inference-Time Scaling with Uncertainty-Aware Tree Search
Directional Textual Inversion for Personalized Text-to-Image Generation
Directly Optimizing Calibrated Test-Time Uncertainty
Dr.LLM: Dynamic Layer Routing in LLMs
DSL-Monkeys: Self-Generated In-Context Examples for Low-Resource GPU DSL Kernels
Efficient Test-Time Adaptation via Decoupled BN Update For Edge Devices
Elbow-based MoE Routing: A Training-Free Inference Time Plugin for Expert Selection
EoRA: Fine-tuning-free Compensation for Compressed LLM with Eigenspace Low-Rank Approximation
EsoLang-Bench: Evaluating Genuine Reasoning in Large Language Models via Esoteric Programming Languages
GradMem: Learning to Write Context into Memory with Test-Time Gradient Descent
HiDRA: A Blazing Fast LM-Head Replacement
Hyper Experts: Language Models With Inference-Time Layer Reallocation
Is Depth Heterogeneity a Barrier to Model Merging?
Joint Consistency: A Unified Test-Time Aggregation Framework with Pairwise Comparisons
KVpop - Retrofitting LLMs with xLSTM-guided Token Eviction
Leveraging RAG for Training-Free Alignment of LLMs
LookSharp: Attention Entropy Minimization for Test-Time Adaptation
Majority Voting For Code Generation
MMT: Achieving Exact Federated Unlearning with Improved Post-Unlearning Performance
MORPHEUS: Meta Test-Time Adaptation via Neural Collapse Geometry
On the Limits of Test-Time Compute: Sequential Reward Filtering for Better Inference
On the Path Dependence of Gradient Ascent-Based Unlearning
Persistent internal state helps maintain learning plasticity
Probing and Steering Chain-of-Thought Unfaithfulness in Language Models
ProtoTTA: Prototype-Guided Test-Time Adaptation
RDUMB++: DRIFT-AWARE CONTINUAL TEST-TIME ADAPTATION
Reference-Guided Machine Unlearning
Refusal-Orthogonal Gated Editing for Safer Localized LLM Adaptation
Reinforced Fast Weights with Next-Sequence Prediction
Rethinking Machine Unlearning: Models Designed to Forget via Key Deletion
Retrieval, Refinement, and Ranking for Text-to-Video Generation via Prompt Optimization and Test-Time Scaling
Self-Improving Vision-Language-Action Models with Data Generation via Residual RL
Self-Soupervision: Cooking and Seasoning Model Soups without Labels for Adaptation
Semantic Anchor Transport: Robust Test-Time Adaptation for Vision-Language Models
SimMerge: Learning to Select Merge Operators from Similarity Signals
Stochastic KV Routing: Enabling Adaptive Depth-Wise Cache Sharing
Terminator: Learning Optimal Exit Points for Early Stopping in Chain-of-Thought Reasoning
Test-Time Adaptation of High-Dimensional Simulation Surrogates via D-Optimal Statistics
Test-time Graph Extrapolation via Progressive Anchor-Guided Expansion
Test-Time Planning for Robust Imitation: Learning to Search for Recovery in Autonomous Driving
Test-Time Self-Distillation
Towards Reasoning Reuse: A New Paradigm in Model Collaboration
TTQ: Activation-Aware Test-Time Quantization to Accelerate LLM Inference On The Fly
When is Model Souping Tasty? Similarity, Transitivity, and Robustness
ZeroSiam: An Efficient Asymmetry for Test-Time Entropy Optimization without Collapse