[ICML 2024] Language Models Represent Beliefs of Self and Others

Visage contains an image dataset of images with human annotations on whether or not certain attributes are present or depicted in the image. The attribute may either be stereotypical or non-stereot…

Code for the paper "Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias"

CompanionLLM - A framework to finetune LLMs to be your own sentient conversational companion

Experimenting with using Knowledge Graphs and AI/LLMs to transform Clinical Practice Guidelines (CPG) into computable Business Process Modeling Notation (BPMN).

A list of awesome papers and resources of recommender system on large language model (LLM).

LLM&ebedding

[NeurIPS 2023] Reflexion: Language Agents with Verbal Reinforcement Learning

A computational approach to the Stereotype Content Model

Open-sourced codes for MiniGPT-4 and MiniGPT-v2 (https://minigpt-4.github.io, https://minigpt-v2.github.io/)

LAVIS - A One-stop Library for Language-Vision Intelligence

1 min voice data can also be used to train a good TTS model! (few shot voice cloning)

A playbook for systematically maximizing the performance of deep learning models.

[NeurIPS 2023] Generalized Logit Adjustment

The official GitHub page for the survey paper "A Survey on Data Augmentation in Large Model Era"

ICLR2024 Spotlight: curation/training code, metadata, distribution and pre-trained models for MetaCLIP; CVPR 2024: MoDE: CLIP Data Experts via Clustering

WordBias: Visualizing Intersectional Social biases encoded in Word Embeddings

The PATA dataset

Code for our ICLR 2024 paper "PerceptionCLIP: Visual Classification by Inferring and Conditioning on Contexts"

[CVPR 2024 Highlight] Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding

PAIRS (PArallel Images for eveRyday Scenarios)

a simple blockchain with pbft

Reconstruction of traditional credit investigation mode based on block chain. (UTXO模型，BOLTDB数据库，PBFT共识算法实现企业征信信息的交换)

[NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.

[ICLR 2024] Fine-tuning LLaMA to follow Instructions within 1 Hour and 1.2M Parameters