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
Unsupervised Language Model Pre-training for French

FlauBERT and FLUE FlauBERT is a French BERT trained on a very large and heterogeneous French corpus. Models of different sizes are trained using the n

GETALP 212 Dec 10, 2022
Training RNNs as Fast as CNNs

News SRU++, a new SRU variant, is released. [tech report] [blog] The experimental code and SRU++ implementation are available on the dev branch which

Tao Lei 14 Dec 12, 2022
Ptorch NLU, a Chinese text classification and sequence annotation toolkit, supports multi class and multi label classification tasks of Chinese long text and short text, and supports sequence annotation tasks such as Chinese named entity recognition, part of speech tagging and word segmentation.

Pytorch-NLU,一个中文文本分类、序列标注工具包,支持中文长文本、短文本的多类、多标签分类任务,支持中文命名实体识别、词性标注、分词等序列标注任务。 Ptorch NLU, a Chinese text classification and sequence annotation toolkit, supports multi class and multi label classifi

186 Dec 24, 2022
Code to use Augmented Shapiro Wilks Stopping, as well as code for the paper "Statistically Signifigant Stopping of Neural Network Training"

This codebase is being actively maintained, please create and issue if you have issues using it Basics All data files are included under losses and ea

Justin Terry 32 Nov 09, 2021
Kurumi ChatBot

KurumiChatBot Just another Telegram AI chat bot written in Python using Pyrogram. A public running instance can be found on telegram as @TokisakiChatB

Yoga Pranata 3 Jun 28, 2022
Neural text generators like the GPT models promise a general-purpose means of manipulating texts.

Boolean Prompting for Neural Text Generators Neural text generators like the GPT models promise a general-purpose means of manipulating texts. These m

Jeffrey M. Binder 20 Jan 09, 2023
Malaya-Speech is a Speech-Toolkit library for bahasa Malaysia, powered by Deep Learning Tensorflow.

Malaya-Speech is a Speech-Toolkit library for bahasa Malaysia, powered by Deep Learning Tensorflow. Documentation Proper documentation is available at

HUSEIN ZOLKEPLI 151 Jan 05, 2023
Code release for "COTR: Correspondence Transformer for Matching Across Images"

COTR: Correspondence Transformer for Matching Across Images This repository contains the inference code for COTR. We plan to release the training code

UBC Computer Vision Group 358 Dec 24, 2022
A deep learning-based translation library built on Huggingface transformers

DL Translate A deep learning-based translation library built on Huggingface transformers and Facebook's mBART-Large 💻 GitHub Repository 📚 Documentat

Xing Han Lu 244 Dec 30, 2022
Applying "Load What You Need: Smaller Versions of Multilingual BERT" to LaBSE

smaller-LaBSE LaBSE(Language-agnostic BERT Sentence Embedding) is a very good method to get sentence embeddings across languages. But it is hard to fi

Jeong Ukjae 13 Sep 02, 2022
xFormers is a modular and field agnostic library to flexibly generate transformer architectures by interoperable and optimized building blocks.

Description xFormers is a modular and field agnostic library to flexibly generate transformer architectures by interoperable and optimized building bl

Facebook Research 2.3k Jan 08, 2023
A Survey of Natural Language Generation in Task-Oriented Dialogue System (TOD): Recent Advances and New Frontiers

A Survey of Natural Language Generation in Task-Oriented Dialogue System (TOD): Recent Advances and New Frontiers

Libo Qin 132 Nov 25, 2022
CJK computer science terms comparison / 中日韓電腦科學術語對照 / 日中韓のコンピュータ科学の用語対照 / 한·중·일 전산학 용어 대조

CJK computer science terms comparison This repository contains the source code of the website. You can see the website from the following link: Englis

Hong Minhee (洪 民憙) 88 Dec 23, 2022
🗣️ NALP is a library that covers Natural Adversarial Language Processing.

NALP: Natural Adversarial Language Processing Welcome to NALP. Have you ever wanted to create natural text from raw sources? If yes, NALP is for you!

Gustavo Rosa 21 Aug 12, 2022
To be a next-generation DL-based phenotype prediction from genome mutations.

Sequence -----------+-- 3D_structure -- 3D_module --+ +-- ? | |

Eric Alcaide 18 Jan 11, 2022
Generate product descriptions, blogs, ads and more using GPT architecture with a single request to TextCortex API a.k.a Hemingwai

TextCortex - HemingwAI Generate product descriptions, blogs, ads and more using GPT architecture with a single request to TextCortex API a.k.a Hemingw

TextCortex AI 27 Nov 28, 2022
Mkdocs + material + cool stuff

Modern-Python-Doc-Example mkdocs + material + cool stuff Doc is live here Features out of the box amazing good looking website thanks to mkdocs.org an

Francesco Saverio Zuppichini 61 Oct 26, 2022
ADCS cert template modification and ACL enumeration

Purpose This tool is designed to aid an operator in modifying ADCS certificate templates so that a created vulnerable state can be leveraged for privi

Fortalice Solutions, LLC 78 Dec 12, 2022
NLP codes implemented with Pytorch (w/o library such as huggingface)

NLP_scratch NLP codes implemented with Pytorch (w/o library such as huggingface) scripts ├── models: Neural Network models ├── data: codes for dataloa

3 Dec 28, 2021