EMNLP 2022: "MABEL: Attenuating Gender Bias using Textual Entailment Data" https://arxiv.org/abs/2210.14975

Code associated with the paper "Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Lists"

Official implementation for KDD'22 paper "Learning Fair Representation via Distributional Contrastive Disentanglement"

BertViz: Visualize Attention in NLP Models (BERT, GPT2, BART, etc.)

This repository contains Pytorch code for AttenD (Attention Debiasing), a finetuning method for reducing social biases from transformer-based text encoders (e.g. BERT, RoBERTa, ALBERT, DistilBert, …

LexGLUE: A Benchmark Dataset for Legal Language Understanding in English

Official repository for "Measuring and Mitigating Gender Bias in Legal Contextualized Language Models".

QLoRA: Efficient Finetuning of Quantized LLMs

This repository contains the data and code introduced in the paper "CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models" (EMNLP 2020).

StereoSet: Measuring stereotypical bias in pretrained language models

This repository contains the required codes for reproducing the results in "Gender Bias in Legal Texts and Debiasing It" which is published in Natural Language Engineering journal.

Fairlex: A Multilingual Benchmark for Evaluating Fairness in Legal Text Processing

Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Control

Learning Gender-Neutral Word Embeddings

A latent text-to-image diffusion model

A collection of resources and papers on Diffusion Models

This is a repo for the EMNLP 19 Paper on gender bias in gendered languages.

replication of Word Embedding Association Test(WEAT), which is suggested in Semantics derived automatically from language corpora necessarily contain human biases by Aylin Caliskan-Islam, Joanna J…

Library for fast text representation and classification.

Cross-lingual version of WEAT

Code and test data for "On Measuring Bias in Sentence Encoders", to appear at NAACL 2019.

Relevant data for the paper Evaluating Bias in Dutch Word Embeddings.

A library for Multilingual Unsupervised or Supervised word Embeddings