# HendrikStrobelt/LSTMVis

Visualization Toolbox for Long Short Term Memory networks (LSTMs)

Repository: https://github.com/HendrikStrobelt/LSTMVis
Canonical: https://ross.abutalabs.com/products/lstmvis
Language: Jupyter Notebook
License: BSD-3-Clause
License Family: permissive
Topics: lstm, neural-network, visualization, recurrent-neural-networks
Last push: 2021-11-19T19:33:59+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3743, "days_push": 1748, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1266, forks 260 (observed 2026-08-28T04:04:11.221035+00:00)

## What it is
LSTMVis is a visual analysis toolbox for inspecting hidden state dynamics of LSTM (and other RNN) networks. It provides a browser-based interactive interface backed by a Python server that consumes HDF5 state files from trained models.

## Use cases
- visualize hidden state changes in an LSTM while processing text
- understand what individual LSTM cells learn during training
- debug and interpret recurrent neural network behavior on sequences
- explore state activations for a children's book or parentheses dataset
- import my own trained model's states into an interactive viewer
- compare RNN hidden states across time steps for interpretability research

## When to choose
- you need interactive, cell-level visual analysis of LSTM/RNN hidden states
- you are doing interpretability research on recurrent networks and can export states to HDF5
- you want a self-hosted web tool with a Python 3.7+ backend

## When to avoid
- you need visualization for transformers or modern architectures rather than RNNs
- you want an actively developed tool with recent updates and community support
- you need a plug-and-play library integrated into your training pipeline rather than a separate server and dataset export step

## Facets
- artifact type: application
- maturity: maintenance
- function: data-visualization, machine-learning, nlp
- domain: machine-learning, data-visualization, deep-learning
- platform: python, browser, cross-platform
- tags: lstm, rnn, hidden-state-analysis, interpretability, d3, hdf5, keras, tensorflow, research-tool, natural-language-processing, web-server

## Member repositories
- HendrikStrobelt/LSTMVis (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:11.221035+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:03:45.562245+00:00, confidence not recorded.
  - readme: https://github.com/HendrikStrobelt/LSTMVis (fetched 2026-08-28T04:04:11.221035+00:00, sha ebc259ad9c08)
- Data as of 2026-08-30T08:39:29.467469+00:00.