diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000000000000000000000000000000000000..c6c3eeb193b462d13cc1c9c97752290f3086b6cf
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,74 @@
+# General
+.pytest_cache
+.DS_Store
+.AppleDouble
+.LSOverride
+__pycache__
+*.pyc
+client_ref/
+_store/
+.cache/
+client_ref
+*.swp
+raw_data
+
+# Icon must end with two \r
+Icon
+
+# Package files
+*.egg-info
+
+# Thumbnails
+._*
+
+# Files that might appear in the root of a volume
+.DocumentRevisions-V100
+.fseventsd
+.Spotlight-V100
+.TemporaryItems
+.Trashes
+.VolumeIcon.icns
+.com.apple.timemachine.donotpresent
+
+# Directories potentially created on remote AFP share
+.AppleDB
+.AppleDesktop
+Network Trash Folder
+Temporary Items
+.apdisk
+
+.idea
+.vscode
+
+.ipynb_checkpoints
+client/src/node_modules
+client/src/.cache-loader
+
+# HDF5 and faiss files can be too big for git
+*.hdf5
+*.faiss
+*.code-workspace
+
+# Datasets
+data/*/wizard-of-oz.txt
+wizard-of-oz.txt
+data/*/wikipedia*.csv.gz
+*_tst
+server/data/*.zip
+allResponses.json
+*.h5py
+*.pckl
+
+# Dev notebooks
+notebooks/*.ipynb
+
+# Demo
+client/src/demo
+
+# Mypy static checking
+.mypy_cache
+.pytest_cache
+
+# Emacs things
+\#*
+.#*
\ No newline at end of file
diff --git a/LICENSE b/LICENSE
new file mode 100644
index 0000000000000000000000000000000000000000..d645695673349e3947e8e5ae42332d0ac3164cd7
--- /dev/null
+++ b/LICENSE
@@ -0,0 +1,202 @@
+
+ Apache License
+ Version 2.0, January 2004
+ http://www.apache.org/licenses/
+
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+ 1. Definitions.
+
+ "License" shall mean the terms and conditions for use, reproduction,
+ and distribution as defined by Sections 1 through 9 of this document.
+
+ "Licensor" shall mean the copyright owner or entity authorized by
+ the copyright owner that is granting the License.
+
+ "Legal Entity" shall mean the union of the acting entity and all
+ other entities that control, are controlled by, or are under common
+ control with that entity. For the purposes of this definition,
+ "control" means (i) the power, direct or indirect, to cause the
+ direction or management of such entity, whether by contract or
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
+ outstanding shares, or (iii) beneficial ownership of such entity.
+
+ "You" (or "Your") shall mean an individual or Legal Entity
+ exercising permissions granted by this License.
+
+ "Source" form shall mean the preferred form for making modifications,
+ including but not limited to software source code, documentation
+ source, and configuration files.
+
+ "Object" form shall mean any form resulting from mechanical
+ transformation or translation of a Source form, including but
+ not limited to compiled object code, generated documentation,
+ and conversions to other media types.
+
+ "Work" shall mean the work of authorship, whether in Source or
+ Object form, made available under the License, as indicated by a
+ copyright notice that is included in or attached to the work
+ (an example is provided in the Appendix below).
+
+ "Derivative Works" shall mean any work, whether in Source or Object
+ form, that is based on (or derived from) the Work and for which the
+ editorial revisions, annotations, elaborations, or other modifications
+ represent, as a whole, an original work of authorship. For the purposes
+ of this License, Derivative Works shall not include works that remain
+ separable from, or merely link (or bind by name) to the interfaces of,
+ the Work and Derivative Works thereof.
+
+ "Contribution" shall mean any work of authorship, including
+ the original version of the Work and any modifications or additions
+ to that Work or Derivative Works thereof, that is intentionally
+ submitted to Licensor for inclusion in the Work by the copyright owner
+ or by an individual or Legal Entity authorized to submit on behalf of
+ the copyright owner. For the purposes of this definition, "submitted"
+ means any form of electronic, verbal, or written communication sent
+ to the Licensor or its representatives, including but not limited to
+ communication on electronic mailing lists, source code control systems,
+ and issue tracking systems that are managed by, or on behalf of, the
+ Licensor for the purpose of discussing and improving the Work, but
+ excluding communication that is conspicuously marked or otherwise
+ designated in writing by the copyright owner as "Not a Contribution."
+
+ "Contributor" shall mean Licensor and any individual or Legal Entity
+ on behalf of whom a Contribution has been received by Licensor and
+ subsequently incorporated within the Work.
+
+ 2. Grant of Copyright License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ copyright license to reproduce, prepare Derivative Works of,
+ publicly display, publicly perform, sublicense, and distribute the
+ Work and such Derivative Works in Source or Object form.
+
+ 3. Grant of Patent License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ (except as stated in this section) patent license to make, have made,
+ use, offer to sell, sell, import, and otherwise transfer the Work,
+ where such license applies only to those patent claims licensable
+ by such Contributor that are necessarily infringed by their
+ Contribution(s) alone or by combination of their Contribution(s)
+ with the Work to which such Contribution(s) was submitted. If You
+ institute patent litigation against any entity (including a
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
+ or a Contribution incorporated within the Work constitutes direct
+ or contributory patent infringement, then any patent licenses
+ granted to You under this License for that Work shall terminate
+ as of the date such litigation is filed.
+
+ 4. Redistribution. You may reproduce and distribute copies of the
+ Work or Derivative Works thereof in any medium, with or without
+ modifications, and in Source or Object form, provided that You
+ meet the following conditions:
+
+ (a) You must give any other recipients of the Work or
+ Derivative Works a copy of this License; and
+
+ (b) You must cause any modified files to carry prominent notices
+ stating that You changed the files; and
+
+ (c) You must retain, in the Source form of any Derivative Works
+ that You distribute, all copyright, patent, trademark, and
+ attribution notices from the Source form of the Work,
+ excluding those notices that do not pertain to any part of
+ the Derivative Works; and
+
+ (d) If the Work includes a "NOTICE" text file as part of its
+ distribution, then any Derivative Works that You distribute must
+ include a readable copy of the attribution notices contained
+ within such NOTICE file, excluding those notices that do not
+ pertain to any part of the Derivative Works, in at least one
+ of the following places: within a NOTICE text file distributed
+ as part of the Derivative Works; within the Source form or
+ documentation, if provided along with the Derivative Works; or,
+ within a display generated by the Derivative Works, if and
+ wherever such third-party notices normally appear. The contents
+ of the NOTICE file are for informational purposes only and
+ do not modify the License. You may add Your own attribution
+ notices within Derivative Works that You distribute, alongside
+ or as an addendum to the NOTICE text from the Work, provided
+ that such additional attribution notices cannot be construed
+ as modifying the License.
+
+ You may add Your own copyright statement to Your modifications and
+ may provide additional or different license terms and conditions
+ for use, reproduction, or distribution of Your modifications, or
+ for any such Derivative Works as a whole, provided Your use,
+ reproduction, and distribution of the Work otherwise complies with
+ the conditions stated in this License.
+
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
+ any Contribution intentionally submitted for inclusion in the Work
+ by You to the Licensor shall be under the terms and conditions of
+ this License, without any additional terms or conditions.
+ Notwithstanding the above, nothing herein shall supersede or modify
+ the terms of any separate license agreement you may have executed
+ with Licensor regarding such Contributions.
+
+ 6. Trademarks. This License does not grant permission to use the trade
+ names, trademarks, service marks, or product names of the Licensor,
+ except as required for reasonable and customary use in describing the
+ origin of the Work and reproducing the content of the NOTICE file.
+
+ 7. Disclaimer of Warranty. Unless required by applicable law or
+ agreed to in writing, Licensor provides the Work (and each
+ Contributor provides its Contributions) on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+ implied, including, without limitation, any warranties or conditions
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+ PARTICULAR PURPOSE. You are solely responsible for determining the
+ appropriateness of using or redistributing the Work and assume any
+ risks associated with Your exercise of permissions under this License.
+
+ 8. Limitation of Liability. In no event and under no legal theory,
+ whether in tort (including negligence), contract, or otherwise,
+ unless required by applicable law (such as deliberate and grossly
+ negligent acts) or agreed to in writing, shall any Contributor be
+ liable to You for damages, including any direct, indirect, special,
+ incidental, or consequential damages of any character arising as a
+ result of this License or out of the use or inability to use the
+ Work (including but not limited to damages for loss of goodwill,
+ work stoppage, computer failure or malfunction, or any and all
+ other commercial damages or losses), even if such Contributor
+ has been advised of the possibility of such damages.
+
+ 9. Accepting Warranty or Additional Liability. While redistributing
+ the Work or Derivative Works thereof, You may choose to offer,
+ and charge a fee for, acceptance of support, warranty, indemnity,
+ or other liability obligations and/or rights consistent with this
+ License. However, in accepting such obligations, You may act only
+ on Your own behalf and on Your sole responsibility, not on behalf
+ of any other Contributor, and only if You agree to indemnify,
+ defend, and hold each Contributor harmless for any liability
+ incurred by, or claims asserted against, such Contributor by reason
+ of your accepting any such warranty or additional liability.
+
+ END OF TERMS AND CONDITIONS
+
+ APPENDIX: How to apply the Apache License to your work.
+
+ To apply the Apache License to your work, attach the following
+ boilerplate notice, with the fields enclosed by brackets "[]"
+ replaced with your own identifying information. (Don't include
+ the brackets!) The text should be enclosed in the appropriate
+ comment syntax for the file format. We also recommend that a
+ file or class name and description of purpose be included on the
+ same "printed page" as the copyright notice for easier
+ identification within third-party archives.
+
+ Copyright [yyyy] [name of copyright owner]
+
+ Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+ You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+ Unless required by applicable law or agreed to in writing, software
+ distributed under the License is distributed on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ See the License for the specific language governing permissions and
+ limitations under the License.
diff --git a/Makefile b/Makefile
new file mode 100644
index 0000000000000000000000000000000000000000..5888cb1e6224570e6ee795024f6493460b7ac9f9
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,8 @@
+ENV_NAME="exbert"
+
+env:
+ conda env create -f ./environment.yml; \
+ conda activate $(ENV_NAME); \
+ pip install -e server/spacyface; \
+ pip install -e server/transformers; \
+ pip install -e server; \
diff --git a/README.md b/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ddf7779fdeb05a47b4f4353d7ade9fa654d61f7b
--- /dev/null
+++ b/README.md
@@ -0,0 +1,217 @@
+# exBERT
+
+[](https://opensource.org/licenses/Apache-2.0)
+
+
+### A Visual Analysis Tool to Explore Learned Representations in Transformers Models
+by Ben Hoover, Hendrik Strobelt, Sebastian Gehrmann
+from IBM Research and Harvard NLP
+
+Link to pre-paper and demo: [exbert.net](http://exbert.net)
+
+
+
+
+
+
+
An overview of the different components of the tool. The token ``escape'' is selected and masked at 0-[all]. The results from a corpus search by token embedding are shown and summarized in (d-g). Users can enter a sentence in (a) and modify the attention view through selections in (b). Self attention is displayed in (c). The blue matrices show the attention of a head (column) to a token (row). Tokens and heads that are selected in (c) can be searched over the annotated corpus (shown: Wizard of Oz) with results presented in (d). Every token in (d) displays its linguistic metadata on hover. A colored summary of the matched token (black highlight) and its context is shown in (e), which can be expanded or collapsed with the buttons above it. The histograms in (f) and (g) summarize the metadata of the results in (d) for the matched token and the token of max attention, respectively.
+
+
+
+### Version 0.9
+
+- [Overview](#overview)
+- [Components](#components)
+- [Install and Getting Started](#install-and-getting-started)
+- [Development](#development)
+
+
+
+
+
+## Overview
+
+exBERT is a tool that enables users to explore what and how transformers learn to model languages. The tool comes equipped with a pretrained base version of BERT, a state of the art architecture based on transformers. To explore what this architecture has learned, you can input any sentence to the model and the tool will parse the sentence into tokens suitable for BERT (using the BPE tokenizer) and pass these tokens through the model. The attentions and ensuing word embeddings of each encoder are then extracted and displayed for interaction.
+
+To ease interpretability of language features, several key features of BERT have been disabled:
+
+1. The attentions toward [CLS] and [SEP] tokens have been zeroed. The [CLS] token is typically useful in classification tasks where the entire sentence needs to have an embedding to summarize it, and the [SEP] has been shown to be a no-op for heads that haven't learned anything.
+2. BERT is able to concatenate two sections of text for training (separated by the [SEP] token mentioned above), and attention patterns can be learned between the two sentences. This enables BERT to apply to a wide range of applications. However, the intention of this tool is to focus on self-attention - that is, the attention of the words in a sentence to other words in the same sentence - and the functionality to look at attention between different words has been dropped.
+
+Even though BERT is able to analyze large chunks of paragraph at once, this tool primarily focuses on language features within one sentence and thus only searches across a corpus that has been split by sentence.
+
+Importantly, BERT's ability to mask particular tokens (by using the [MASK] token in place of the original token) has been preserved.
+
+## Components
+
+
+
+
+### 1) The Attention Explorer
+
+
+#### 1a) Sentence Input
+
+Type in any english sentence you want to analyze in the ensuing visualization. Just note that the longer the sentence, the longer the visualization will require to initialize.
+
+
+#### 1b) Threshold Slider
+
+The Threshold Slider is used to control how much attention is displayed in the main attention graph. For each word, show the largest attentions until X amount of the total attention has been displayed. At 1, all the attention connections are shown. At 0, nothing is shown. Use this if you are only interested in seeing what each word is most interested in.
+
+
+#### 1c) Layer Selection
+
+Choose the layer of BERT to analyze
+
+
+#### 1d) Head Selector
+
+Display the selected heads, with the option to select all or none. For convenience, interaction has been added to the Attention Summary Boxes that allows you to select or deselect heads from the visualization itself.
+
+
+#### 1e) Attention Summary Boxes
+
+Every row represents the total attention to a token, and every column represents a different head. By looking down a column, you can see how strongly any particular head is activated at a particular head/layer. By hovering over any column, you can see that attentions that belong to just that head in the main graph. By selecting a column, you can select or deselect that head. This will both remove the effect of that head on the overall visualization and indicate which heads you are interested in searching over in the corpus.
+
+The left side indicates how much total attention each head is going out of each word. The right side indicates the total attention from each head going into each word.
+
+
+#### 1f) Attention Graph
+
+This central display shows how tokens attend to other tokens in the same sentence. By hovering over any token, you can see only the attentions going into or out of that particular word. By single clicking on it, you can freeze the view of that particular token and explore how the heads interact with only that token's attention. This also indicates which embedding / headlist you are interested in searching the corpus for. By double clicking on the token, you can mask that token, which passes all the tokens back to BERT with the [MASK] token replacing the token you just double clicked. This often changes quite a few of the attentions and will rerender the attention graph. You can then continue to explore the attentions and select tokens and head as usual.
+
+
+### 2) Corpus Explorer
+
+Right now, the only available corpus to search is the Wizard of Oz (WoZ). This corpus has been split into sentences, parsed for language features such as part of speech (POS), dependency (DEP), and entity information using SPACY, merged into the BERT tokenization scheme, stored into an HDF5 file, and indexed by FAISS for quick lookup. This same procedure will need to be applied to other corpora to be searched.
+
+
+#### 2a) Search Buttons:
+
+There are two buttons to perform a search of the available corpus: Search by Head and Search by Embedding.
+
+To search by embedding, the embedding of the selected token and the layer at which to compare embeddings are passed to the backend. The processed corpus is searched for the words that are nearest to the embedding (by inner product search) at that particular layer and displayed in the Corpus Explorer.
+
+To search by head, the selected layer, the selected heads, and the concatenated head vector for the selected token are passed to the backend. The head vector is then set to 0 at all indices that are from one of the unselected heads. This vector is then searched across the corpus by inner product.
+
+Note that the term "inner product" is used. Since the embeddings and the head vectors are normalized, this is equivalent to performing a cosine-similarity search.
+
+
+#### 2b) Histogram Information
+
+There are two different histograms of information that are displayed: the Metadata histogram (in purple) and the positional histogram (in black). By selecting the dropdown underneath the Metadata histogram, you can change the displayed matrix in the Corpus controller.
+
+
+#### 2c) Corpus Metadata Matrix control buttons
+
+The control buttons allow you to see a certain amount of context on either side of the matched word. The arrows add context one word to the left or right, whereas the red X deletes a context from the left or right. The blue refresh symbol is used to adjust the heights of each cell to correspond to the height of it's sentence, which is important when the browser window has been resized in some way.
+
+
+#### 2d) Corpus Metadata Matrix
+
+The metadata matrix is an array of colors that summarize the metadata information of the corresponding sentence to its right. By hovering over any cell, you will be able to see what that particular color represents. There are unfortunately too many values for POS and DEP to give each a unique color that is distinct from all other colors, so some colors may overlap. Black cells indicate that you have reached a sentence boundary.
+
+
+#### 2e) Corpus Explorer
+
+This display shows all words most closely matching the selected token / layer / head information indicated in the Attention Explorer. Matched words have a thick red border. Hovering over any word will give you its POS and DEP information, the amount of attention the matched word is paying to that word, and will read ENTITY if that word was determined to be an entity in the original corpus.
+
+## Install and Getting Started
+
+Note: This code has possible OS dependencies as it was developed exclusively on MacOS.
+
+### Setting up the Environment
+1. From the root of this project, create a new conda directory with `conda env create -f environment.yml`. This will create an environment named `exbert`.
+2. Activate this environment with `conda activate exbert`. At this point, if you want to install the development dependencies, you can do so with `conda env update -f environment-dev.yml`
+3. You will need to install spacy's `en_core_web_sm` as well. To do this, run: `python -m spacy download en_core_web_sm`
+
+
+### Generating Example Data
+
+please see the [instructions here](https://github.com/bhoov/exbert/tree/master/server/data_processing)
+
+
+### Running Locally
+Starting the backend:
+
+```bash
+conda activate exbert
+python server/main.py
+```
+
+### Notes on setting up conda
+If setting up conda for the first time, we recommend downloading Miniconda with the following curl command:
+
+```
+curl 'https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh'
+```
+
+Promptly refresh your shell environment and run `conda update conda` to be able to install from the `conda env create` command above.
+
+
+## Development
+
+If you want to make custom changes to the code, these are some hints to get you started.
+
+### Use as package
+Some find it useful to expose the code inside `server` for development in an environment like Jupyter Notebooks. From the root folder with the `exbert` environment active:
+
+```bash
+conda env update -f environment-dev.yml
+pip install -e ./server
+```
+
+Now the `exbert` environment should allow the server code to be accessible in any folder so long as there are no additional module name clashes in the environment.
+
+### Compiling the frontend
+
+```bash
+cd client/src
+npm install #installs all necessary node packages
+npm run build #This will create the static files living in `client/dist`.
+```
+
+## Running a development environment
+You can run a client server that automatically recompiles the frontend with `npm run ww`. After making a change, you should be able to refresh the browser window to see your most recent changes.
+
+Because the backend has to load in a lot of data for inference, we do not allow auto-backend refresh on every saved change in flask even though the framework supports it.
+
+### Uploading your own model locally
+Uploading your own model consists of the following steps:
+
+1. *Save your pretrained huggingface model* according to the naming conventions specified in the "modeling_auto.py" of the original transformers repo:
+
+```
+The model class to instantiate is selected as the first pattern matching
+ in the `pretrained_model_name_or_path` string (in the following order):
+ - contains `t5`: T5Model (T5 model)
+ - contains `distilbert`: DistilBertModel (DistilBERT model)
+ - contains `albert`: AlbertModel (ALBERT model)
+ - contains `camembert`: CamembertModel (CamemBERT model)
+ - contains `xlm-roberta`: XLMRobertaModel (XLM-RoBERTa model)
+ - contains `roberta`: RobertaModel (RoBERTa model)
+ - contains `bert`: BertModel (Bert model)
+ - contains `openai-gpt`: OpenAIGPTModel (OpenAI GPT model)
+ - contains `gpt2`: GPT2Model (OpenAI GPT-2 model)
+ - contains `transfo-xl`: TransfoXLModel (Transformer-XL model)
+ - contains `xlnet`: XLNetModel (XLNet model)
+ - contains `xlm`: XLMModel (XLM model)
+ - contains `ctrl`: CTRLModel (Salesforce CTRL model)
+```
+
+Right now, only BERT, RoBERTa, GPT2, and DistilBERT are supported for context searching. You can use the rest without the context searching as desired.
+
+2. *Create the reference corpus*. **Warning**: Depending on the number of layers and size of the hidden dimension in the model, this step could take many gigabytes on your computer to store the hidden representations and attentions at every layer.
+
+## Notes on SubRepo Usage
+This project makes use of two public pip repositories (`transformers` and `spacyface`), both of which needed modification as this project was being developed. The `git-subrepo` tool was used to achieve this workflow with a forked repository of both transformers and spacyface. However, this introduces the following steps when setting up the environment:
+
+1. From the `transformers/` directory, run `pip install -e .`
+2. Repeat for the `spacyface/` directory.
+
+## Acknowledgements
+This project was inspired in part by the original [BertViz by Jesse Vig](https://github.com/jessevig/bertviz).
+
+## Debugging
+- If you get a `No module named '_swigfaiss'` error, check that `libomp` is installed on your system. If you are on a mac, this is as simple as `brew install libomp`.
diff --git a/client/dist/05ca9c06114e79436ea9b5c8d4a7869c.ttf b/client/dist/05ca9c06114e79436ea9b5c8d4a7869c.ttf
new file mode 100644
index 0000000000000000000000000000000000000000..4a890111a095b61af17e2ca4239254a6347803d6
Binary files /dev/null and b/client/dist/05ca9c06114e79436ea9b5c8d4a7869c.ttf differ
diff --git a/client/dist/4171e41154ba857f85c536f167d581ba.ttf b/client/dist/4171e41154ba857f85c536f167d581ba.ttf
new file mode 100644
index 0000000000000000000000000000000000000000..031c5ad54de2af6b5a3950b923f2bd0fb2463c4f
Binary files /dev/null and b/client/dist/4171e41154ba857f85c536f167d581ba.ttf differ
diff --git a/client/dist/7eeb10384e8e1ef96c87f7074cf2ef59.ttf b/client/dist/7eeb10384e8e1ef96c87f7074cf2ef59.ttf
new file mode 100644
index 0000000000000000000000000000000000000000..ce42e55bafd813315a35dbb44fd0564ad6e95a20
Binary files /dev/null and b/client/dist/7eeb10384e8e1ef96c87f7074cf2ef59.ttf differ
diff --git a/client/dist/a849e7649e2005ab4aecfa50d96120e1.ttf b/client/dist/a849e7649e2005ab4aecfa50d96120e1.ttf
new file mode 100644
index 0000000000000000000000000000000000000000..4b9c15f0134a75362a474749aa6e1878b1df87cd
Binary files /dev/null and b/client/dist/a849e7649e2005ab4aecfa50d96120e1.ttf differ
diff --git a/client/dist/exBERT.html b/client/dist/exBERT.html
new file mode 100644
index 0000000000000000000000000000000000000000..081db10e82a7598e4a33a4c406635de2594eab6d
--- /dev/null
+++ b/client/dist/exBERT.html
@@ -0,0 +1,226 @@
+
+
+
+
+
+
+
+
+ exBERT
+
+
+
+
+
+
+
+
+
+
+
+ An Explorable BERT
+
+
+
+ IBM Research & HarvardNLP
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Layer:
+
+
+
+
+
+
+ Hide Special Tokens
+
+
+
+
+
+
+
+
+
+ Selected heads:
+
+
+
+
+
+
+
+
+
+
+
You focus on one token by click.
+ You can mask any token by double click.
+
You can select and de-select a head by a click on the heatmap columns
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Search
+
+
+
+
+
+
+
+
+
+
+
+ Matched Word Summary:
+
+
+
+
+
+
+
+
+
+
+
+ Max Attention Summary:
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ ←||→
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/client/dist/img/annotated_instructions2.png b/client/dist/img/annotated_instructions2.png
new file mode 100644
index 0000000000000000000000000000000000000000..dc5f9b966e2782404b314f0b7aa81cd570498523
Binary files /dev/null and b/client/dist/img/annotated_instructions2.png differ
diff --git a/client/dist/img/exBERT.png b/client/dist/img/exBERT.png
new file mode 100644
index 0000000000000000000000000000000000000000..c17abeb6127a67055063b0d1c8c0b781b6ef0bb3
Binary files /dev/null and b/client/dist/img/exBERT.png differ
diff --git a/client/dist/img/exBERT.svg b/client/dist/img/exBERT.svg
new file mode 100644
index 0000000000000000000000000000000000000000..52bcc8c1077f25792d463d0c3064c2d35ae1b850
--- /dev/null
+++ b/client/dist/img/exBERT.svg
@@ -0,0 +1,12 @@
+
+
+
diff --git a/client/dist/img/exBERT_favicon.png b/client/dist/img/exBERT_favicon.png
new file mode 100644
index 0000000000000000000000000000000000000000..d1f5d36ce1637f06da8f7656bdd62316852c63d2
Binary files /dev/null and b/client/dist/img/exBERT_favicon.png differ
diff --git a/client/dist/img/exBERT_overview.png b/client/dist/img/exBERT_overview.png
new file mode 100644
index 0000000000000000000000000000000000000000..c8d99cfa71ced20abbbaa0cc9451966e21bd5141
Binary files /dev/null and b/client/dist/img/exBERT_overview.png differ
diff --git a/client/dist/img/exbert_teaser_V2.png b/client/dist/img/exbert_teaser_V2.png
new file mode 100644
index 0000000000000000000000000000000000000000..7f843c6e48fb796f23f2aa0d05b8c4adc43873b8
Binary files /dev/null and b/client/dist/img/exbert_teaser_V2.png differ
diff --git a/client/dist/img/favicon.png b/client/dist/img/favicon.png
new file mode 100644
index 0000000000000000000000000000000000000000..7285a46f9dc7edcd6259c8207c3ebcf0800c20bd
Binary files /dev/null and b/client/dist/img/favicon.png differ
diff --git a/client/dist/img/harvardnlp_logo.png b/client/dist/img/harvardnlp_logo.png
new file mode 100644
index 0000000000000000000000000000000000000000..eb57f712b25cebf46174e22e998ab8b393476f57
Binary files /dev/null and b/client/dist/img/harvardnlp_logo.png differ
diff --git a/client/dist/index.html b/client/dist/index.html
new file mode 100644
index 0000000000000000000000000000000000000000..03774870d641820df4af9e12ac9d8a20bd8d4380
--- /dev/null
+++ b/client/dist/index.html
@@ -0,0 +1,236 @@
+
+
+
+
+
+
+
+
+ exBERT
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ Large language models can produce powerful contextual representations that lead to improvements across many
+ NLP tasks.
+ Since these models are typically guided by a sequence of learned self attention mechanisms and may comprise
+ undesired inductive biases, it is paramount to be able to explore what the attention has learned.
+ While static analyses of these models lead to targeted insights, interactive tools are more dynamic and can
+ help humans better gain an intuition for the model-internal reasoning process.
+
+
+
+ We present exBERT , an interactive tool named after the popular BERT language model, that provides
+ insights into the meaning of the contextual representations by matching a human-specified input to similar
+ contexts in a large annotated dataset.
+ By aggregating the annotations of the matching similar contexts, exBERT helps intuitively explain
+ what each attention-head has learned.
+
+
+
Large language models can produce powerful contextual representations that lead to improvements across many
+ NLP tasks. Though these models can comprise undesired inductive biases, it is challenging to identify what
+ information they encode in their learned representations.
+
+
Since the model-internal reasoning process is often guided by a sequence of learned self-attention
+ mechanisms, it is paramount to be able to explore what the attention has learned. While static analyses for
+ this can lead to targeted insights, interactive tools can be more dynamic and help humans gain an intuition
+ for the model-internal reasoning process. We present exBERT, a tool that helps to gain insights into the
+ meaning of the contextual representations. exBERT matches a human-specified input to similar contexts in a
+ large annotated dataset. By aggregating these annotations across all similar contexts, exBERT can help to
+ explain what each attention-head has learned.
+
+
Thanks to
+ Jesse Vig
+ for feedback. Please let us know what you think by commenting below!
+
+
+ 