Skip to content
Papers.

Learning to Discover at Test Time research paper by Stanford University, 2026

Stanford University · Jan 22, 2026 · Retrieval and data · 82 citations · 45 upvotes · unverified

Read on arXiv

What it shows

Test-time training enables AI systems to discover optimal solutions for specific scientific problems through continual learning focused on individual challenges rather than generalization.

UnverifiedHugging Face's summary; not yet checked by hand.

More from Stanford University

All 9
PaperCitations
FAMOS: Feed-Forward 3D Articulation Modeling from Sparse ObservationsModeling articulated objects from sparse monocular views is challenging because each observation reveals only partial geometry and motion evidence.Inference and efficiency · Sep 2026 · Unverified0
AISPA: User-Centric System Prompt Auditing for Large Language Model ApplicationsSystem prompts are instructions configured by developers to govern the behaviors of foundation models in AI applications.Foundation models · Jul 2026 · Unverified0
Recursive Multi-Agent SystemsRecursiveMAS extends recursive scaling principles from single models to multi-agent systems, enabling collaborative reasoning through iterative latent-space computations with improved efficiency and accuracy.Agents and evaluation · Apr 2026 · Unverified6
Multi-User Large Language Model AgentsMulti-user large language model agents face challenges in handling conflicting objectives, privacy preservation, and coordination efficiency in multi-principal decision-making scenarios.Foundation models · Mar 2026 · Unverified2
In-the-Flow Agentic System Optimization for Effective Planning and Tool UseAgentFlow, a trainable agentic framework with in-the-flow optimization, enhances reasoning in large language models by coordinating specialized modules and outperforms top baselines across various tasks.Agents and evaluation · Oct 2025 · Unverified58
LongCodeZip: Compress Long Context for Code Language ModelsLongCodeZip is a code compression framework for LLMs that uses dual-stage compression to reduce context size without degrading performance, improving efficiency in code intelligence applications.Inference and efficiency · Oct 2025 · Unverified47
Direct Preference Optimization: Your Language Model is Secretly a Reward ModelDPO aligns a model to human preferences with a simple classification loss, no reward model or RL loop.Alignment and safety · May 202310.6k
FlashAttention: Fast and Memory-Efficient Exact Attention with IO-AwarenessFlashAttention computes exact attention with far fewer GPU memory reads and writes, making long sequences faster.Inference and efficiency · May 20225,334

Other retrieval and data papers

Topic
PaperCitations
From Local to Global: A Graph RAG Approach to Query-Focused SummarizationGraphRAG builds a knowledge graph of a document set so a model can answer questions about the whole collection.Microsoft · Apr 20242,198
DINOv3DINOv3, a self-supervised learning model, achieves superior performance across various vision tasks by scaling datasets and models, addressing dense feature degradation, and enhancing flexibility with post-hoc strategies.Meta · Aug 2025 · Unverified1,498
DeepSeek-Prover: Advancing Theorem Proving in LLMs through Large-Scale Synthetic DataGenerating extensive Lean 4 proof data from mathematical competition problems improved DeepSeekMath 7B's theorem-proving capabilities over GPT-4 and other methods.DeepSeek · May 2024 · Unverified260
Qwen3-VL-Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the-Art Multimodal Retrieval and RankingThe Qwen3-VL-Embedding and Qwen3-VL-Reranker models form an end-to-end multimodal search pipeline, leveraging multi-stage training and cross-attention mechanisms to achieve high-precision retrieval across diverse modalities.Alibaba (Qwen) · Jan 2026 · Unverified231
DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree SearchDeepSeek-Prover-V1.5 improves theorem proving by optimizing training and inference, utilizing reinforcement learning, and proposing RMaxTS for diverse proof paths, achieving state-of-the-art results on miniF2F and ProofNet benchmarks.DeepSeek · Aug 2024 · Unverified207
Arctic-Embed: Scalable, Efficient, and Accurate Text Embedding ModelsOpen text embedding models from 22M to 334M parameters that led MTEB retrieval for their size at release.Snowflake · May 202480
About this paper
Authors
Mert Yuksekgonul, Daniel Koceja, Xinhao Li and 8 more
arXiv
2601.16175 · PDF
Venue
arXiv.org
Citations
82, 10 influential · Semantic Scholar
Upvotes
45 · Hugging Face
Code
github.com/test-time-training/discover
Lab
Stanford University

Changes

What changed
Influential citationsfirst count: 10Sep 25, 2026
Citationsfirst count: 82Sep 25, 2026
New paperFound by the weekly scan, unverifiedSep 25, 2026

Sources: each lab's own papers and arXiv, with citation and upvote counts from Semantic Scholar and Hugging Face. One-line summaries are for orientation, not a substitute for the paper. Logos via logo.dev; trademarks belong to their owners.

New papers by email

Monday afternoons, only in weeks with new papers from the labs.

Double opt-in. Unsubscribe any time.