Physics of Language Models: Part 4.2, Canon Layers at Scale where Synthetic Pretraining Resonates in Reality
☆322Jan 5, 2026Updated last month
Alternatives and similar repositories for PhysicsLM4
Users that are interested in PhysicsLM4 are comparing it to the libraries listed below
Sorting:
- ☆129Jun 6, 2025Updated 8 months ago
- [ICLR 2025 & COLM 2025] Official PyTorch implementation of the Forgetting Transformer and Adaptive Computation Pruning☆140Updated this week
- Revisiting Mid-training in the Era of Reinforcement Learning Scaling☆183Jul 23, 2025Updated 7 months ago
- The code for creating the iGSM datasets in papers "Physics of Language Models Part 2.1, Grade-School Math and the Hidden Reasoning Proces…☆84Jan 12, 2025Updated last year
- ☆156Updated this week
- 🔥 A minimal training framework for scaling FLA models☆352Nov 15, 2025Updated 3 months ago
- Measuring the Signal to Noise Ratio in Language Model Evaluation☆28Aug 19, 2025Updated 6 months ago
- Organize the Web: Constructing Domains Enhances Pre-Training Data Curation☆78May 2, 2025Updated 10 months ago
- [ICLR 2025] Official PyTorch Implementation of Gated Delta Networks: Improving Mamba2 with Delta Rule☆477Feb 17, 2026Updated 2 weeks ago
- ☆32Feb 11, 2025Updated last year
- 🚀 Efficient implementations of state-of-the-art linear attention models☆4,428Updated this week
- Stick-breaking attention☆62Jul 1, 2025Updated 8 months ago
- Muon is Scalable for LLM Training☆1,440Aug 3, 2025Updated 7 months ago
- Code for NeurIPS 2024 Spotlight: "Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations"☆92Oct 30, 2024Updated last year
- ☆122Feb 4, 2026Updated 3 weeks ago
- Official codebase for our paper "Do Language Models Use Their Depth Efficiently?"☆29Jun 25, 2025Updated 8 months ago
- [ICLR 26] The official code repository for the paper "Mirage or Method? How Model–Task Alignment Induces Divergent RL Conclusions".☆15Feb 9, 2026Updated 3 weeks ago
- The official repository for SkyLadder: Better and Faster Pretraining via Context Window Scheduling☆42Dec 29, 2025Updated 2 months ago
- The official implementation for [NeurIPS2025 Oral] Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink…☆861Dec 20, 2025Updated 2 months ago
- Open-source framework for the research and development of foundation models.☆773Updated this week
- ☆99Aug 8, 2025Updated 6 months ago
- Code release for "Understanding Bias in Large-Scale Visual Datasets"☆22Dec 4, 2024Updated last year
- Open-sourcing code associated with the AAAI-25 paper "On the Expressiveness and Length Generalization of Selective State-Space Models on …☆14Sep 18, 2025Updated 5 months ago
- The simplest implementation of recent Sparse Attention patterns for efficient LLM inference.☆91Jul 17, 2025Updated 7 months ago
- Skill-It! A Data-Driven Skills Framework for Understanding and Training Language Models☆48Oct 31, 2023Updated 2 years ago
- Supporting code for the blog post on modular manifolds.☆117Sep 26, 2025Updated 5 months ago
- Code for paper "Patch-Level Training for Large Language Models"☆96Nov 10, 2025Updated 3 months ago
- Code for ICML 25 paper "Metadata Conditioning Accelerates Language Model Pre-training (MeCo)"☆50Jun 30, 2025Updated 8 months ago
- Here we will test various linear attention designs.☆62Apr 25, 2024Updated last year
- Homepage for ProLong (Princeton long-context language models) and paper "How to Train Long-Context Language Models (Effectively)"☆247Sep 12, 2025Updated 5 months ago
- ☆63Jun 12, 2025Updated 8 months ago
- Memory layers use a trainable key-value lookup mechanism to add extra parameters to a model without increasing FLOPs. Conceptually, spars…☆372Dec 12, 2024Updated last year
- The official github repo for "Diffusion Language Models are Super Data Learners".☆223Nov 6, 2025Updated 3 months ago
- Code release for paper "Test-Time Training Done Right"☆379Jan 5, 2026Updated last month
- Official Repo for Error-Free Linear Attention is a Free Lunch: Exact Solution from Continuous-Time Dynamics☆71Jan 13, 2026Updated last month
- ☆50Jun 16, 2025Updated 8 months ago
- Best practices for training DeepSeek, Mixtral, Qwen and other MoE models using Megatron Core.☆166Jan 22, 2026Updated last month
- ☆451Aug 10, 2025Updated 6 months ago
- Mamba support for transformer lens☆19Sep 17, 2024Updated last year