PaulPauls / llama3_interpretability_sae
A complete end-to-end pipeline for LLM interpretability with sparse autoencoders (SAEs) using Llama 3.2, written in pure PyTorch and fully reproducible.
☆461Updated this week
Related projects ⓘ
Alternatives and complementary repositories for llama3_interpretability_sae
- Bayesian Optimization as a Coverage Tool for Evaluating LLMs. Accurate evaluation (benchmarking) that's 10 times faster with just a few l…☆271Updated this week
- Revealing example of self-attention, the building block of transformer AI models☆130Updated last year
- ☆162Updated 5 months ago
- a curated list of data for reasoning ai☆113Updated 3 months ago
- DiscoGrad - automatically differentiate across conditional branches in C++ programs☆204Updated 2 months ago
- Visualize the intermediate output of Mistral 7B☆316Updated 9 months ago
- Implement recursion using English as the programming language and an LLM as the runtime.☆128Updated last year
- A copy of ONNX models, datasets, and code all in one GitHub repository. Follow the README to learn more.☆106Updated last year
- An implementation of bucketMul LLM inference☆214Updated 4 months ago
- This project collects GPU benchmarks from various cloud providers and compares them to fixed per token costs. Use our tool for efficient …☆210Updated last month
- Visual inference exploration & experimentation playground☆78Updated this week
- ai for jq☆234Updated 2 months ago
- Visualizing the internal board state of a GPT trained on chess PGN strings, and performing interventions on its internal board state and …☆193Updated this week
- Run and explore Llama models locally with minimal dependencies on CPU☆183Updated last month
- Stop messing around with finicky sampling parameters and just use DRµGS!☆318Updated 5 months ago
- Mistral7B playing DOOM☆122Updated 4 months ago
- A Detailed Introduction to My Favorite Statistical Measure, Hoeffding's D☆95Updated 9 months ago
- ☆223Updated last month
- Official codebase for the paper "Beyond A* Better Planning with Transformers via Search Dynamics Bootstrapping".☆322Updated 5 months ago
- Radient turns many data types (not just text) into vectors for similarity search, RAG, regression analysis, and more.☆269Updated 2 months ago
- Grow virtual creatures in static and physics simulated environments.☆52Updated 8 months ago
- Absolute minimalistic implementation of a GPT-like transformer using only numpy (<650 lines).☆250Updated last year
- Action library for AI Agent☆191Updated this week
- The Fast Vector Similarity Library is designed to provide efficient computation of various similarity measures between vectors.☆351Updated 2 months ago
- Text generator prompting with Boolean operators☆180Updated last year
- Command-line interface for the Arcane Engine☆43Updated 3 weeks ago
- Extensible AI assistant platform that bridges LLMs to tasks and actions☆38Updated last year
- What if an HNSW index was just a file, and you could serve it from a CDN, and search it directly in the browser?☆86Updated 6 months ago
- Array-Inspired Pipeline Language☆119Updated last year
- Agent Based Model on GPU using CUDA 12.2.1 and OpenGL 4.5 (CUDA OpenGL interop) on Windows/Linux☆69Updated last month