dynamical-inference / patchsae
Implementation of PatchSAE as presented in "Sparse autoencoders reveal selective remapping of visual concepts during adaptation". Code coming soon.
☆10Updated 3 months ago
Alternatives and similar repositories for patchsae:
Users that are interested in patchsae are comparing it to the libraries listed below
- ☆20Updated 4 months ago
- Official pytorch implementation of "Interpreting the Second-Order Effects of Neurons in CLIP"☆34Updated 4 months ago
- ☆35Updated 6 months ago
- Official PyTorch Implementation for Task Vectors are Cross-Modal☆22Updated 3 months ago
- What do we learn from inverting CLIP models?☆53Updated last year
- Unsolvable Problem Detection: Evaluating Trustworthiness of Vision Language Models☆75Updated 6 months ago
- Implementation of PaCE: Parsimonious Concept Engineering for Large Language Models (NeurIPS 2024)☆35Updated 4 months ago
- An official implementation of "Catastrophic Failure of LLM Unlearning via Quantization" (ICLR 2025)☆26Updated last month
- ☆10Updated 5 months ago
- ☆48Updated 4 months ago
- PyTorch implementation for our paper "Improving GFlowNets for Text-to-Image Diffusion Alignment."☆23Updated 6 months ago
- ☆11Updated 10 months ago
- This is the repository for "Model Merging by Uncertainty-Based Gradient Matching", ICLR 2024.☆27Updated 10 months ago
- Official repo for Detecting, Explaining, and Mitigating Memorization in Diffusion Models (ICLR 2024)☆69Updated 11 months ago
- ☆17Updated 2 weeks ago
- Official implementation of MAIA, A Multimodal Automated Interpretability Agent☆76Updated 3 weeks ago
- ☆13Updated last year
- ☆31Updated 2 months ago
- Intriguing Properties of Data Attribution on Diffusion Models (ICLR 2024)☆28Updated last year
- [NAACL 2025 Oral] Multimodal Needle in a Haystack (MMNeedle): Benchmarking Long-Context Capability of Multimodal Large Language Models☆40Updated 3 weeks ago
- ☆28Updated 3 years ago
- Official implementation of the paper The Hidden Language of Diffusion Models☆72Updated last year
- Auto Interpretation Pipeline and many other functionalities for Multimodal SAE Analysis.☆124Updated 2 months ago
- Erasing conceptual knowledge from language models through low-rank fine-tuning☆12Updated this week
- ☆28Updated 8 months ago
- ☆11Updated 2 months ago
- ☆40Updated 8 months ago
- ☆21Updated 10 months ago
- Model Merging with SVD to Tie the KnOTS [ICLR 2025]☆45Updated 2 months ago
- Official Pytorch implementation of "Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations" (ICLR '25)☆63Updated 3 weeks ago