Tweet fijado
Inseq
57 posts
Inseq
@InseqLib
Open-Source Interpretability for Generative Language Models 🔎 🐛
Europe Se unió Kasım 2022
611 Siguiendo342 Seguidores
Inseq retuiteado
Model Internals-based RAG Evaluation (MIRAGE) 🌴 is accepted to #EMNLP2024 Main!
➡️ To celebrate, here's our new MIRAGE demo combining @InseqLib and Transformers-specific LRP: huggingface.co/spaces/gsarti/….
Reach out if you want to catch up in Miami! 🤗🏖️
Jirui Qi@Jirui_Qi
[1/8] Struggling with verifying the trustworthiness of RAG outputs? Check our latest work where we utilize *model internals* as a powerful and faithful tool for attributing answers to retrieved docs! (w/ @gsarti_ @AriannaBisazza @raquel_dmg) 📄: arxiv.org/abs/2406.13663 #NLProc
English
Inseq retuiteado
Very hyped for the new beautiful viz that just landed in the @InseqLib main branch! 🔥 This will empower users to explore attribution tensors more flexibly and intuitively.
h/t to @_ddjohnson for his awesome work on the treescope toolkit powering this release!
Inseq@InseqLib
Thanks to the new treescope integration, @InseqLib now supports interactive visualizations for multidimensional attributions (show_granular), token highlights (show_tokens) and improved viz for attribute_context CLI! 🚀 Install main, will appear in v0.7 x.com/_ddjohnson/sta…
English
Thanks to the new treescope integration, @InseqLib now supports interactive visualizations for multidimensional attributions (show_granular), token highlights (show_tokens) and improved viz for attribute_context CLI! 🚀 Install main, will appear in v0.7
x.com/_ddjohnson/sta…
Daniel Johnson@_ddjohnson
By popular demand, the Treescope pretty-printer from the Penzai neural net library can now be installed separately, and supports both JAX and PyTorch! And that's not all: Penzai itself now has less boilerplate and includes more pretrained Transformer models!
English
Inseq retuiteado
🎓 We're thrilled to host Gabriele Sarti (@gsarti_) in our PhD seminar series tomorrow, July 16th, from 12:00-13:00 in Oe67 BU 101! Join us for his talk on interpreting context usage in generative language models, featuring the Inseq toolkit and PECoRe framework. 🕛Don't miss it!
English
The 🐑 PECoRe / 🌴 MIRAGE demo on @huggingface Spaces is powered by our new attribute-context CLI command released in v0.6, and allows to export the code to reproduce your results locally with 🐛 Inseq.
Check it out ➡️ hf.co/spaces/gsarti/…
Gabriele Sarti@gsarti_
⚠️ Citations from prompting or NLI seem plausible, but may not faithfully reflect LLM reasoning. 🏝️ MIRAGE detects context dependence in generations via model internals, producing granular and faithful RAG citations. 🚀 Demo: huggingface.co/spaces/gsarti/… Fun collab w/ @Jirui_Qi, @AriannaBisazza & @raquel_dmg! Check it out ⬇️
English
Inseq retuiteado
Today, we had the first seminar of our #XAI course!
@gsarti_ presented the @InseqLib to interpret LMs and the PECORE framework to identify & attribute context dependence in LMs! 🚀🌟
Thank you, it was so interesting! 🤗
Great start to our series!
gsarti.com/talk/polito-in…
English
Inseq retuiteado
[1/4] Introducing “A Primer on the Inner Workings of Transformer-based Language Models”, a comprehensive survey on interpretability methods and the findings into the functioning of language models they have led to.
ArXiv: arxiv.org/pdf/2405.00208
English
@InseqLib What an amazing image to celebrate that number! 🤣
Congrats!
English
@InseqLib v0.6 is out now on PyPI! 🔥
New CLI command for context attribution (@gsarti_), new perturbation-based methods by @hmohebbi75
& @casszzx and optimizations incl. multi-gpu support! ⚡️
Huge shoutout to our contributors! ❤️
Release notes ⬇️
github.com/inseq-team/ins…
English
Want to learn how to detect and attribute context usage in LMs using @InseqLib? This @Gradio demo using our CLI can teach you how! 🐛
Gabriele Sarti@gsarti_
The official 🐑 PECoRe 🐑 demo to detect and attribute context dependence in LM generations is now available on @huggingface Spaces! 🚀 Includes code examples, a usage guide, useful presets for various dec-only & enc-dec models, and more! Check it out ⬇️ huggingface.co/spaces/gsarti/…
English
Inseq retuiteado
🐑 PECoRe repository is now public (github.com/gsarti/pecore) and all model/datasets are available on @huggingface (huggingface.co/collections/gs…)!
🐑Interested in using PECoRe on your models? Have a look at the @InseqLib implementation (`inseq attribute-context`)!
Gabriele Sarti@gsarti_
[1/8] Our new work (w/ @AriannaBisazza @gchrupala @MalvinaNissim) is finally out! 🎉 We introduce PECoRe, an interpretability framework using model internals to identify & attribute context dependence in language models. 📄Paper: arxiv.org/abs/2310.01188 #NLProc #neuralempty
English
Tracker for Mamba gradient-based methods support: github.com/huggingface/tr…
English
Did someone say Mamba support? 👀
(Actually gradient-based methods are not working out-of-the-box because of some in-place variable overwriting, but should be fixable!)
Arthur Zucker@art_zucker
`mamba` is now available in transformers. PEFT finetuning example: gist.github.com/ArthurZucker/7… Thanks @_albertgu and @tri_dao for this brilliant model! 🚀 and the amazing `mamba-ssm` kernels powering this!
English
Inseq retuiteado
Excited to present my recent work on @InseqLib and the #ICLR2024 PECoRe interpretability framework for the @SheffieldNLP group this afternoon! Many thanks @casszzx for inviting me! 🤗
🐛 Inseq: github.com/inseq-team/ins…
🐑🐑 PECoRe: openreview.net/forum?id=XTHfN…
English
Inseq retuiteado
Wow!! This is great news🥳
Thanks @InseqLib! I’m excited to try it out!
Inseq@InseqLib
Value Zeroing, a faithful approach for analyzing context mixing in Transformers, is now available on @InseqLib main branch for all @huggingface text generation models! 🔀 🔍Paper introducing VZ: aclanthology.org/2023.eacl-main… 🐛VZ in Inseq: tinyurl.com/inseq-vz
English
Value Zeroing, a faithful approach for analyzing context mixing in Transformers, is now available on @InseqLib main branch for all @huggingface text generation models! 🔀
🔍Paper introducing VZ: aclanthology.org/2023.eacl-main…
🐛VZ in Inseq: tinyurl.com/inseq-vz
English
After two years with Poetry, Inseq just moved to @astral_sh's new blazing fast package manager uv! Our CI installation step is now ~80% faster! 🔥
Congrats to @charliermarsh and the team on the release, and godspeed for the cargo-like experience you are planning to build!
English
Interested in using @InseqLib for your work? Refer to our new tutorial to get started ➡️ github.com/inseq-team/ins… 8/8
English
Wang et al. propose a metric for LMs factual reliability and show its relation to models’ sensitivity to in-context distractors arxiv.org/abs/2310.09820 (W. Wang @bazril @alexandrabirch1 W. Peng @EdinburghNLP) 7/
English
Mousavi et al. attribute the generation of dialogue models, finding increased influence for more refined versions of dialogue history aclanthology.org/2023.nlp4conva… (@mahedmousavi S. Caldarella G. Riccardi @sislab7) 3/
English