Attribution-based Parameter Decomposition
☆33Jun 11, 2025Updated 8 months ago
Alternatives and similar repositories for apd
Users that are interested in apd are comparing it to the libraries listed below
Sorting:
- Codebase for "On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback". This repo implements a generative multi-tur…☆23Dec 3, 2024Updated last year
- ☆150Dec 30, 2025Updated last month
- Open source replication of Anthropic's Crosscoders for Model Diffing☆64Oct 27, 2024Updated last year
- A benchmark for mechanistic discovery of circuits in Transformers☆16Dec 15, 2024Updated last year
- Engine for collecting, uploading, and downloading model activations☆26Apr 2, 2025Updated 10 months ago
- Stochastic Parameter Decomposition☆65Feb 20, 2026Updated last week
- ☆13Oct 5, 2025Updated 4 months ago
- ☆17Updated this week
- A library for efficient patching and automatic circuit discovery.☆90Dec 31, 2025Updated 2 months ago
- ☆395Aug 21, 2025Updated 6 months ago
- Materials for "Multi-property Steering of Large Language Models with Dynamic Activation Composition"☆14Nov 22, 2024Updated last year
- Code for the "Overcoming Sparsity Artifacts in Crosscoders to Interpret Chat-Tuning" paper.☆16Nov 21, 2025Updated 3 months ago
- Exploring the Limitations of Large Language Models on Multi-Hop Queries☆32Mar 2, 2025Updated 11 months ago
- Code for reproducing our paper "Not All Language Model Features Are Linear"☆84Nov 27, 2024Updated last year
- Code to reproduce key results accompanying "SAEs (usually) Transfer Between Base and Chat Models"☆13Jul 18, 2024Updated last year
- Providing the answer to "How to do patching on all available SAEs on GPT-2?". It is an official repository of the implementation of the p…☆13Jan 26, 2025Updated last year
- ☆20Nov 15, 2024Updated last year
- A tiny easily hackable implementation of a feature dashboard.☆15Oct 21, 2025Updated 4 months ago
- ☆17Aug 30, 2025Updated 6 months ago
- The nnsight package enables interpreting and manipulating the internals of deep learned models.☆825Updated this week
- ☆70Mar 6, 2025Updated 11 months ago
- CS194-196 Course Project☆14Feb 20, 2025Updated last year
- Tools for studying developmental interpretability in neural networks.☆125Dec 30, 2025Updated 2 months ago
- Code for my NeurIPS 2024 ATTRIB paper titled "Attribution Patching Outperforms Automated Circuit Discovery"☆47May 31, 2024Updated last year
- ☆66Feb 16, 2023Updated 3 years ago
- ☆23Jun 30, 2025Updated 8 months ago
- Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.' This library lets language models …☆243Updated this week
- Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).☆245Dec 16, 2024Updated last year
- Code for "On Measuring Faithfulness of Natural Language Explanations"☆21Jul 23, 2024Updated last year
- A library for mechanistic anomaly detection☆22Jan 9, 2025Updated last year
- Code & data for EMNLP 2020 paper "MOCHA: A Dataset for Training and Evaluating Reading Comprehension Metrics".☆16May 3, 2022Updated 3 years ago
- Minimum Description Length probing for neural network representations☆20Jan 28, 2025Updated last year
- Zotero plugin/addon to automatically include the name of the currently selected collection as a tag to a new item.☆19Sep 10, 2024Updated last year
- ☆83Feb 25, 2025Updated last year
- Code for our paper "Decomposing The Dark Matter of Sparse Autoencoders"☆23Feb 6, 2025Updated last year
- Official codebase for "Quantile Reward Policy Optimization: Alignment with Pointwise Regression and Exact Partition Functions" (Matrenok …☆30Dec 8, 2025Updated 2 months ago
- Evaluate interpretability methods on localizing and disentangling concepts in LLMs.☆57Oct 30, 2025Updated 3 months ago
- Landing page for MIB: A Mechanistic Interpretability Benchmark☆24Aug 15, 2025Updated 6 months ago
- [NAACL 2022] GlobEnc: Quantifying Global Token Attribution by Incorporating the Whole Encoder Layer in Transformers☆21May 16, 2023Updated 2 years ago