inseq-team/inseq
An attribution-based interpretability library for sequence generation models including LLMs.
Not currently ranked — collecting fresh signals.
star history
Inseq is an open-source library providing feature attribution methods to explain sequence generation models. It integrates with Hugging Face Transformers and Captum, offering tools to identify which input tokens most influenced model outputs during generation. The library supports various attribution techniques for analyzing LLM behavior in tasks like translation, summarization, and text generation.
Frequently asked
- What is inseq-team/inseq?
- An attribution-based interpretability library for sequence generation models including LLMs.
- Is inseq open source?
- Yes — inseq-team/inseq is open source, released under the Apache-2.0 license.
- What language is inseq written in?
- inseq-team/inseq is primarily written in Python.
- How popular is inseq?
- inseq-team/inseq has 469 stars on GitHub.
- Where can I find inseq?
- inseq-team/inseq is on GitHub at https://github.com/inseq-team/inseq.