Papers of the day   All papers

The Language Interpretability Tool: Extensible, Interactive Visualizations and Analysis for NLP Models

Comments

Ian Tenney: Introducing an early-release version of the Language Interpretability Tool (LIT), a visual, interactive, and extensible open-source tool for analyzing all sorts of NLP models 🔥 Code: https://github.com/pair-code/lit/ Paper: https://arxiv.org/abs/2008.05122 #NLProc (1/4) https://t.co/YooNvmbmQ6

4 replies, 553 likes


Jasmijn Bastings: We built 🔥 LIT! Extensible, Interactive Visualizations and Analysis for NLP Models. More info, links and video below!

2 replies, 221 likes


Fernanda Viégas: Excited to see the Language Interpretability Tool (LIT) launched! Data visualization for NLP models; what could be better? :) This is a collaboration between #googlePAIR and NLP researchers at @googIeresearch

1 replies, 33 likes


Sebastian Gehrmann: If you are working on NLP interpretability, try LIT 🔥 We are looking for all kinds of feedback on our early release. Props to @iftenney and @bengiswex for spearheading this huge effort.

0 replies, 29 likes


James Wexler: Check out our Language Interpretability Tool (LIT), a collab between @GoogleAI PAIR and Language teams. I'm really proud of this work and so excited to see where it goes from here. Consider it a super-powered What-If Tool for NLP models 🔥 #NLProc

0 replies, 26 likes


Andreas Madsen: Woa, amazing too see more Interpretability ideas making their way into NLP.

0 replies, 19 likes


Andy Coenen: NLP might be the most exciting domain in ML, and is likely the field where interpretability is most critical. LIT is a huge step towards that goal. I had an absolute blast working with some brilliant colleagues on this project. Check it out!

2 replies, 16 likes


Dipanjan Das (@🏠): Check out LIT, a tool for interpreting models of natural language. This was co-led by close colleague @iftenney from the Mountain View part of @GoogleAI's Language team.

0 replies, 10 likes


Nikhil Thorat: Language Interpretability Tool (LIT): a tool / UI which allows probing language models! It lets you query counterfactuals, look at attention vectors, view embedding projections, and more. LIT is built by some of my coworkers on PAIR, this is a personal favorite project!

1 replies, 7 likes


arxiv: The Language Interpretability Tool: Extensible, Interactive Visualizations and Analysis f... http://arxiv.org/abs/2008.05122 https://t.co/m12nogE07W

0 replies, 4 likes


Mahima Pushkarna: 🆕📢 The early release of the Language Interpretability Tool is out and it's 🔥 #LIT ! Give us your suggestions and feedback by opening an issue on our GitHub repo <-- we're watching this space closely! 👀 #MachineLearning #NLP #uxresearch #uxdesign

0 replies, 2 likes


Francesco Ronzano: LIT, interactive python-based tool to explore the Interpretability of NLP (Language) Models / Code https://github.com/PAIR-code/lit / Paper https://arxiv.org/abs/2008.05122 / "Why did my model make this prediction? When does it perform poorly? What happens under a controlled change in the input?"

0 replies, 1 likes


arXiv in review: #EMNLP2020 The Language Interpretability Tool: Extensible, Interactive Visualizations and Analysis for NLP Models. (arXiv:2008.05122v1 [cs\.CL]) http://arxiv.org/abs/2008.05122

0 replies, 1 likes


Content

Found on Aug 18 2020 at https://arxiv.org/pdf/2008.05122.pdf

PDF content of a computer science paper: The Language Interpretability Tool: Extensible, Interactive Visualizations and Analysis for NLP Models