2025

an archive of posts from this year

Dec 26, 2025 SliderSpace: Decomposing the Visual Capabilities of Diffusion Models
Dec 26, 2025 SliderSpace: Decomposing the Visual Capabilities of Diffusion Models
Dec 25, 2025 Generative Modeling by Estimating Gradients of the Data Distribution
Dec 25, 2025 What are Diffusion Models?
Dec 24, 2025 Building Diffusion Model's theory from ground up
Dec 07, 2025 Weight-sparse transformers have interpretable circuits
Dec 07, 2025 Between the Bars: Gradient-based Jailbreaks are Bugs that induce Features
Dec 07, 2025 Breakpoint: Scalable evaluation of system-level reasoning in LLM code agents
Dec 06, 2025 Fluid Language Model Benchmarking
Dec 05, 2025 Auditing language models for hidden objectives
Dec 04, 2025 DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning
Dec 03, 2025 DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
Nov 30, 2025 Multiagent Finetuning: Self Improvement with Diverse Reasoning Chains
Nov 24, 2025 ML Tea: Planning and Problem-Solving with General, Scalable Neuro-Symbolic Models
Nov 24, 2025 Automated Detection of Visual Attribute Reliance with a Self-Reflective Agent
Nov 23, 2025 Continuous Language Model Interpolation yields Dynamic and Controllable Text Generation
Nov 23, 2025 Natural Emergent Misalignment from Reward Hacking in Production RL
Nov 22, 2025 A Mathematical Framework for Transformer Circuits
Nov 19, 2025 Value Augmented Sampling for Language Model Alignment and Personalization
Nov 18, 2025 Deriving Muon
Nov 18, 2025 Curiosity-driven Red-teaming for Large Language Models
Nov 18, 2025 Guided Speculative Inference for Efficient Test-Time Alignment of LLMs
Nov 18, 2025 Superhuman AI for Stratego Using Self-Play Reinforcement Learning and Test-Time Search
Nov 17, 2025 Domain-Aware Scaling Laws Uncover Data Synergy
Nov 17, 2025 Data Debiasing with Datamodels (D3M): Improving Subgroup Robustness via Data Selection
Nov 17, 2025 Ambient Diffusion Omni: Training Good Models with Bad Data
Nov 16, 2025 Boomerang Distillation Enables Zero-Shot Model Size Interpolation
Nov 15, 2025 To Backtrack or Not to Backtrack: When Sequential Search Limits Model Reasoning
Nov 15, 2025 Training Language Models to Self-Correct via Reinforcement Learning
Nov 14, 2025 ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
Nov 13, 2025 Large Language Diffusion Models
Nov 12, 2025 Teaching AI to see the world more like we do
Nov 11, 2025 Self-Adapting Language Models
Nov 11, 2025 The Surprising Effectiveness of Test-Time Training for Few-Shot Learning
Nov 11, 2025 Reasoning or reciting? Exploring the capabilities and limitations of language models through counterfactual tasks
Nov 08, 2025 Nested Learning: The Illusion of Deep Learning Architecture
Nov 04, 2025 LoRA Without Regret