A library for finding knowledge neurons in pretrained transformer models.

Overview

knowledge-neurons

An open source repository replicating the 2021 paper Knowledge Neurons in Pretrained Transformers by Dai et al., and extending the technique to autoregressive models, as well as MLMs.

The Huggingface Transformers library is used as the backend, so any model you want to probe must be implemented there.

Currently integrated models:

BERT_MODELS = ["bert-base-uncased", "bert-base-multilingual-uncased"]
GPT2_MODELS = ["gpt2"]
GPT_NEO_MODELS = [
    "EleutherAI/gpt-neo-125M",
    "EleutherAI/gpt-neo-1.3B",
    "EleutherAI/gpt-neo-2.7B",
]

The technique from Dai et al. has been used to locate knowledge neurons in the huggingface bert-base-uncased model for all the head/relation/tail entities in the PARAREL dataset. Both the neurons, and more detailed results of the experiment are published at bert_base_uncased_neurons/*.json and can be replicated by running pararel_evaluate.py. More details in the Evaluations on the PARAREL dataset section.

Setup

Either clone the github, and run scripts from there:

git clone knowledge-neurons
cd knowledge-neurons

Or install as a pip package:

pip install knowledge-neurons

Usage & Examples

An example using bert-base-uncased:

from knowledge_neurons import KnowledgeNeurons, initialize_model_and_tokenizer, model_type
import random

# first initialize some hyperparameters
MODEL_NAME = "bert-base-uncased"

# to find the knowledge neurons, we need the same 'facts' expressed in multiple different ways, and a ground truth
TEXTS = [
    "Sarah was visiting [MASK], the capital of france",
    "The capital of france is [MASK]",
    "[MASK] is the capital of france",
    "France's capital [MASK] is a hotspot for romantic vacations",
    "The eiffel tower is situated in [MASK]",
    "[MASK] is the most populous city in france",
    "[MASK], france's capital, is one of the most popular tourist destinations in the world",
]
TEXT = TEXTS[0]
GROUND_TRUTH = "paris"

# these are some hyperparameters for the integrated gradients step
BATCH_SIZE = 20
STEPS = 20 # number of steps in the integrated grad calculation
ADAPTIVE_THRESHOLD = 0.3 # in the paper, they find the threshold value `t` by multiplying the max attribution score by some float - this is that float.
P = 0.5 # the threshold for the sharing percentage

# setup model & tokenizer
model, tokenizer = initialize_model_and_tokenizer(MODEL_NAME)

# initialize the knowledge neuron wrapper with your model, tokenizer and a string expressing the type of your model ('gpt2' / 'gpt_neo' / 'bert')
kn = KnowledgeNeurons(model, tokenizer, model_type=model_type(MODEL_NAME))

# use the integrated gradients technique to find some refined neurons for your set of prompts
refined_neurons = kn.get_refined_neurons(
    TEXTS,
    GROUND_TRUTH,
    p=P,
    batch_size=BATCH_SIZE,
    steps=STEPS,
    coarse_adaptive_threshold=ADAPTIVE_THRESHOLD,
)

# suppress the activations at the refined neurons + test the effect on a relevant prompt
# 'results_dict' is a dictionary containing the probability of the ground truth being generated before + after modification, as well as other info
# 'unpatch_fn' is a function you can use to undo the activation suppression in the model. 
# By default, the suppression is removed at the end of any function that applies a patch, but you can set 'undo_modification=False', 
# run your own experiments with the activations / weights still modified, then run 'unpatch_fn' to undo the modifications
results_dict, unpatch_fn = kn.suppress_knowledge(
    TEXT, GROUND_TRUTH, refined_neurons
)

# suppress the activations at the refined neurons + test the effect on an unrelated prompt
results_dict, unpatch_fn = kn.suppress_knowledge(
    "[MASK] is the official language of the solomon islands",
    "english",
    refined_neurons,
)

# enhance the activations at the refined neurons + test the effect on a relevant prompt
results_dict, unpatch_fn = kn.enhance_knowledge(TEXT, GROUND_TRUTH, refined_neurons)

# erase the weights of the output ff layer at the refined neurons (replacing them with zeros) + test the effect
results_dict, unpatch_fn = kn.erase_knowledge(
    TEXT, refined_neurons, target=GROUND_TRUTH, erase_value="zero"
)

# erase the weights of the output ff layer at the refined neurons (replacing them with an unk token) + test the effect
results_dict, unpatch_fn = kn.erase_knowledge(
    TEXT, refined_neurons, target=GROUND_TRUTH, erase_value="unk"
)

# edit the weights of the output ff layer at the refined neurons (replacing them with the word embedding of 'target') + test the effect
# we can make the model think the capital of france is London!
results_dict, unpatch_fn = kn.edit_knowledge(
    TEXT, target="london", neurons=refined_neurons
)

for bert models, the position where the "[MASK]" token is located is used to evaluate the knowledge neurons, (and the ground truth should be what the mask token is expected to be), but due to the nature of GPT models, the last position in the prompt is used by default, and the ground truth is expected to immediately follow.

In GPT models, due to the subword tokenization, the integrated gradients are taken n times, where n is the length of the expected ground truth in tokens, and the mean of the integrated gradients at each step is taken.

for bert models, the ground truth is currently expected to be a single token. Multi-token ground truths are on the todo list.

Evaluations on the PARAREL dataset

To ensure that the repo works correctly, figures 3 and 4 from the knowledge neurons paper are reproduced below. In general the results appear similar, except suppressing unrelated facts appears to have a little more of an affect in this repo than in the paper's original results.*

Below are Dai et al's, and our result, respectively, for suppressing the activations of the refined knowledge neurons in pararel: knowledge neuron suppression / dai et al. knowledge neuron suppression / ours

And Dai et al's, and our result, respectively, for enhancing the activations of the knowledge neurons: knowledge neuron enhancement / dai et al. knowledge neuron enhancement / ours

To find the knowledge neurons in bert-base-uncased for the PARAREL dataset, and replicate figures 3. and 4. from the paper, you can run

# find knowledge neurons + test suppression / enhancement (this will take a day or so on a decent gpu) 
# you can skip this step since the results are provided in `bert_base_uncased_neurons`
python -m torch.distributed.launch --nproc_per_node=NUM_GPUS_YOU_HAVE pararel_evaluate.py
# plot results 
python plot_pararel_results.py

*It's unclear where the difference comes from, but my suspicion is they made sure to only select facts with different relations, whereas in the plots below, only a different pararel UUID was selected. In retrospect, this could actually express the same fact, so I'll rerun these experiments soon.

TODO:

  • Better documentation
  • Publish PARAREL results for bert-base-multilingual-uncased
  • Publish PARAREL results for bert-large-uncased
  • Publish PARAREL results for bert-large-multilingual-uncased
  • Multiple masked tokens for bert models
  • Find good dataset for GPT-like models to evaluate knowledge neurons (PARAREL isn't applicable since the tail entities aren't always at the end of the sentence)
  • Add negative examples for getting refined neurons (i.e expressing a different fact in the same way)
  • Look into different attribution methods (cf. https://arxiv.org/pdf/2010.02695.pdf)

Citations

@article{Dai2021KnowledgeNI,
  title={Knowledge Neurons in Pretrained Transformers},
  author={Damai Dai and Li Dong and Y. Hao and Zhifang Sui and Furu Wei},
  journal={ArXiv},
  year={2021},
  volume={abs/2104.08696}
}
Owner
EleutherAI
EleutherAI
Learning Energy-Based Models by Diffusion Recovery Likelihood

Learning Energy-Based Models by Diffusion Recovery Likelihood Ruiqi Gao, Yang Song, Ben Poole, Ying Nian Wu, Diederik P. Kingma Paper: https://arxiv.o

Ruiqi Gao 41 Nov 22, 2022
Codebase for arXiv preprint "NeRF++: Analyzing and Improving Neural Radiance Fields"

NeRF++ Codebase for arXiv preprint "NeRF++: Analyzing and Improving Neural Radiance Fields" Work with 360 capture of large-scale unbounded scenes. Sup

Kai Zhang 722 Dec 28, 2022
Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network

Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network Paddle-PANet 目录 结果对比 论文介绍 快速安装 结果对比 CTW1500 Method Backbone Fine

7 Aug 08, 2022
A benchmark dataset for emulating atmospheric radiative transfer in weather and climate models with machine learning (NeurIPS 2021 Datasets and Benchmarks Track)

ClimART - A Benchmark Dataset for Emulating Atmospheric Radiative Transfer in Weather and Climate Models Official PyTorch Implementation Using deep le

21 Dec 31, 2022
Official PyTorch implementation of the paper "TEMOS: Generating diverse human motions from textual descriptions"

TEMOS: TExt to MOtionS Generating diverse human motions from textual descriptions Description Official PyTorch implementation of the paper "TEMOS: Gen

Mathis Petrovich 187 Dec 27, 2022
CLDF dataset derived from Robbeets et al.'s "Triangulation Supports Agricultural Spread" from 2021

CLDF dataset derived from Robbeets et al.'s "Triangulation Supports Agricultural Spread" from 2021 How to cite If you use these data please cite the o

Digital Linguistics 2 Dec 20, 2021
App customer segmentation cohort rfm clustering

CUSTOMER SEGMENTATION COHORT RFM CLUSTERING TỔNG QUAN VỀ HỆ THỐNG DỮ LIỆU Nên chuyển qua theme màu dark thì sẽ nhìn đẹp hơn https://customer-segmentat

hieulmsc 3 Dec 18, 2021
EM-POSE 3D Human Pose Estimation from Sparse Electromagnetic Trackers.

EM-POSE: 3D Human Pose Estimation from Sparse Electromagnetic Trackers This repository contains the code to our paper published at ICCV 2021. For ques

Facebook Research 62 Dec 14, 2022
Code for "Learning the Best Pooling Strategy for Visual Semantic Embedding", CVPR 2021

Learning the Best Pooling Strategy for Visual Semantic Embedding Official PyTorch implementation of the paper Learning the Best Pooling Strategy for V

Jiacheng Chen 106 Jan 06, 2023
SoK: Vehicle Orientation Representations for Deep Rotation Estimation

SoK: Vehicle Orientation Representations for Deep Rotation Estimation Raymond H. Tu, Siyuan Peng, Valdimir Leung, Richard Gao, Jerry Lan This is the o

FIRE Capital One Machine Learning of the University of Maryland 12 Oct 07, 2022
PyTorch-centric library for evaluating and enhancing the robustness of AI technologies

Responsible AI Toolbox A library that provides high-quality, PyTorch-centric tools for evaluating and enhancing both the robustness and the explainabi

24 Dec 22, 2022
Megaverse is a new 3D simulation platform for reinforcement learning and embodied AI research

Megaverse Megaverse is a new 3D simulation platform for reinforcement learning and embodied AI research. The efficient design of the engine enables ph

Aleksei Petrenko 191 Dec 23, 2022
classification task on dataset-CIFAR10,by using Tensorflow/keras

CIFAR10-Tensorflow classification task on dataset-CIFAR10,by using Tensorflow/keras 在这一个库中,我使用Tensorflow与keras框架搭建了几个卷积神经网络模型,针对CIFAR10数据集进行了训练与测试。分别使

3 Oct 17, 2021
An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicity.

Fast Face Classification (F²C) This is the code of our paper An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicit

33 Jun 27, 2021
Invert and perturb GAN images for test-time ensembling

GAN Ensembling Project Page | Paper | Bibtex Ensembling with Deep Generative Views. Lucy Chai, Jun-Yan Zhu, Eli Shechtman, Phillip Isola, Richard Zhan

Lucy Chai 93 Dec 08, 2022
Hide screen when boss is approaching.

BossSensor Hide your screen when your boss is approaching. Demo The boss stands up. He is approaching. When he is approaching, the program fetches fac

Hiroki Nakayama 6.2k Jan 07, 2023
Applications using the GTN library and code to reproduce experiments in "Differentiable Weighted Finite-State Transducers"

gtn_applications An applications library using GTN. Current examples include: Offline handwriting recognition Automatic speech recognition Installing

Facebook Research 68 Dec 29, 2022
LiDAR R-CNN: An Efficient and Universal 3D Object Detector

LiDAR R-CNN: An Efficient and Universal 3D Object Detector Introduction This is the official code of LiDAR R-CNN: An Efficient and Universal 3D Object

TuSimple 295 Jan 05, 2023
A note taker for NVDA. Allows the user to create, edit, view, manage and export notes to different formats.

Quick Notetaker add-on for NVDA The Quick Notetaker add-on is a wonderful tool which allows writing notes quickly and easily anytime and from any app

5 Dec 06, 2022
Implementation of Perceiver, General Perception with Iterative Attention in TensorFlow

Perceiver This Python package implements Perceiver: General Perception with Iterative Attention by Andrew Jaegle in TensorFlow. This model builds on t

Rishit Dagli 84 Oct 15, 2022