Fake10086
Follow
Stars
Resources for skilling up in AI alignment research engineering. Covers basics of deep learning, mechanistic interpretability, and RL.
ViT Prisma is a mechanistic interpretability library for Vision Transformers (ViTs).
This is the first released survey paper on hallucinations of large vision-language models (LVLMs). To keep track of this field and continuously update our survey, we maintain this repository of rel…
[ICML 2024] Memory-Space Visual Prompting for Efficient Vision-Language Fine-Tuning
LLaVA-UHD: an LMM Perceiving Any Aspect Ratio and High-Resolution Images
Efficient Multimodal Large Language Models: A Survey
[ECCV 2024] Code for paper: An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models
A comprehensive survey on Internal Consistency and Self-Feedback in Large Language Models, including theoretical frameworks, task classifications, evaluation methods, future research directions and…
Knowledge Graphs Meet Multi-Modal Learning: A Comprehensive Survey
Easy multi-task learning with HuggingFace Datasets and Trainer
The trainer for HF to record losses of different tasks and objectives.
Repository for Label Words are Anchors: An Information Flow Perspective for Understanding In-Context Learning
Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".
Training Sparse Autoencoders on Language Models
Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).
[CVPR 2024] Prompt Highlighter: Interactive Control for Multi-Modal LLMs
The accompanying code for "Transformer Feed-Forward Layers Are Key-Value Memories". Mor Geva, Roei Schuster, Jonathan Berant, and Omer Levy. EMNLP, 2021.
Creative interactive views of any dataset.
The attention heads in the Transformer architecture possess a variety of capabilities. This is a carefully compiled list that summarizes the diverse functions of the attention heads.
My solutions to DLFC - Deep Learning: Foundations and Concepts
This repository contains the code for running the character-level Sandwich Transformers from our ACL 2020 paper on Improving Transformer Models by Reordering their Sublayers.