torchlens

Inspect, visualize and intervene on any operation in any neural network. Capture every activation and gradient, the full computational graph and 800+ metadata fields without changing model code. Supports PyTorch, JAX, TensorFlow, MLX, PaddlePaddle and tinygrad. Validated on 6400+ models. Built for researchers and their AI agents.

johnmarktaylor91/torchlens

View on GitHub ↗

Inspect, visualize and intervene on any operation in any neural network. Capture every activation and gradient, the full computational graph and 800+ metadata fields without changing model code. Supports PyTorch, JAX, TensorFlow, MLX, PaddlePaddle and tinygrad. Validated on 6400+ models. Built for researchers and their AI agents.

torchlens ranks 10th of 80 in Interpretability & Alignment, actively maintained, last commit 5 days ago.

Star trajectory

16 days tracked
2026-11-06714 stars today2026-11-22
Momentum

- in 30d, below the Interpretability & Alignment median of +0.4%.

Maintenance

Last commit 5 days ago, fresher than the median of 183d.

How it compares in Interpretability & Alignment

torchlens
714
stars
- 30d
commit 5 days ago
Interpretability & Alignment median
290
stars
+0.4% 30d
commit ~183d ago
Leader · lmms-eval
4.5k
stars
+0.9% 30d
commit today

Ranked 10th of 80 by stars.

Ranked by shared tags, then keywords, then stars

Where torchlens sits

Hover or focus a node for details. Click to open it.

  • This repo
  • Related repo
  • Tag or category
  1. Graph convolutional neural network for multirelational link prediction

    +0 (+0.0%) in 30d · Committed 1,461d ago · Jupyter Notebook · MIT

    Shares Interpretability & Alignment

  2. A PyTorch implementation of ACM SIGKDD 2019 paper "Predicting Dynamic Embedding Trajectory in Temporal…

    +0 (+0.0%) in 30d · Committed 849d ago · Python · MIT

    Shares Interpretability & Alignment

  3. The purpose of this repository is to make prototypes as case study in the context of proof of concept(PoC) and…

    +0 (+0.0%) in 30d · Committed 1,062d ago · Python · GPL-2.0

    Shares Interpretability & Alignment

  4. Refusal removal as a loadable KV-cache graft — zero weight modification, no refusal-direction projection…

    +93 (+51%) in 30d · Committed 61d ago · Python · MIT

    Shares Interpretability & Alignment

  5. An Integrated Library for Tuning, Deploying and Interpreting Genomic Models

    +1 (+0.8%) in 30d · Committed 230d ago · Python

    Shares Interpretability & Alignment

  6. Auditing LLM-as-judge systems with paired synthetic controls, bootstrap CIs, and BH-FDR bias detection.

    -2 (-3.7%) in 30d · Committed 86d ago · Python · MIT

    Shares Interpretability & Alignment

  7. A JAX research toolkit for building, editing, and visualizing neural networks.

    -2 (-0.1%) in 30d · Committed 518d ago · Python · Apache-2.0

    Shares Interpretability & Alignment

  8. Observe how memory shapes an answer. Pensieve is an interactive system for visualizing, interpreting, and…

    -6 (-0.5%) in 30d · Committed 66d ago · TypeScript · MIT

    Shares Interpretability & Alignment

  9. A small, fast decision model, or system one model, for agentic workflows. Pick between options or rate on a…

    +550 (+79%) in 30d · Committed yesterday · Python · Apache-2.0

    Shares Interpretability & Alignment

  10. [ACL2023] We introduce LLM-Blender, an innovative ensembling framework to attain consistently superior…

    +2 (+0.2%) in 30d · Committed 760d ago · Python · Apache-2.0

    Shares Interpretability & Alignment

Language
Python
License
Apache License 2.0
Contributors
9
Forks
35
Watchers
5
Open issues
0
Releases
132
Created
2022-10-14
Stars this week
+0

You've seen where torchlens stands today.

Get the moves before everyone else: the 5 Interpretability & Alignment repos climbing fastest, every Tuesday.

Free weekly digest. Double opt-in, unsubscribe anytime.

Maintain torchlens?

Claim this page to add your own call-to-action.

Claim this page