The nnsight package enables interpreting and manipulating the internals of deep learned models.

tree is a library for working with nested data structures

Density Estimation from Approximate Nearest Neighbors

Home repository for the ACORN dataset: 3,500 explanations with aspect-wise human ratings of their quality.

A byte-level decoder architecture that matches the performance of tokenized Transformers.

Analyzing Cognitive Plausibility of Subword Tokenization

A comprehensive benchmark for entity disambiguation

AKGR: Awesome Knowledge Graph Reasoning is a collection of knowledge graph reasoning works, including papers, codes and datasets

Are foundation LMs multilingual knowledge bases? (EMNLP 2023)

The hub for EleutherAI's work on interpretability and learning dynamics

PaCMAP: Large-scale Dimension Reduction Technique Preserving Both Global and Local Structure

GLADIS: A General and Large Acronym Disambiguation Benchmark (EACL 23)

Evaluating German T5 Models on GermEval 2014 (NER)

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

ExtremeBERT is a toolkit that accelerates the pretraining of customized language models on customized datasets, described in the paper “ExtremeBERT: A Toolkit for Accelerating Pretraining of Custom…

Cross-stitch bi-encoder for distantly supervised relation extraction (EMNLP 2022)

Language model fine-tuning on NER with an easy interface and cross-domain evaluation. "T-NER: An All-Round Python Library for Transformer-based Named Entity Recognition, EACL 2021"

Zoomable, animated scatterplots in the browser that scales over a billion points

CoDEx: A set of knowledge graph Completion Datasets Extracted from Wikidata and Wikipedia

μKG: A Library for Multi-source Knowledge Graph Embeddings and Applications, ISWC 2022

Differentiable Algorithms and Algorithmic Supervision.

Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions

A library for finding knowledge neurons in pretrained transformer models.

Code for Comparing Text Representations: A Theory-Driven Approach

🤖 A Python library for learning and evaluating knowledge graph embeddings

Code for the RecAdam paper: Recall and Learn: Fine-tuning Deep Pretrained Language Models with Less Forgetting.

GC4LM: A Colossal (Biased) language model for German

[ACL 20] TACRED Revisited: A Thorough Evaluation of the TACRED Relation Extraction Task

Minimum-distortion embedding with PyTorch