TokenPrint: open-source interactive explorer for tokens, attention, hidden states and KV cache seeks contributors
Rich-Fruit-326 · reddit · 2026-09-10
TokenPrint is an open-source project to explore transformer/LLM internals interactively — tokens, embeddings, attention, hidden states, inference steps, KV cache, and architecture — beyond static diagrams.
It has reached 65+ GitHub stars with active issues, and the author is recruiting contributors: good first issues, advanced Python/PyTorch/Transformers work, React/Three.js visualization, GGUF/llama.cpp and local LLMs, UI/UX, and new ideas or criticism. Repo: Sudharsanselvaraj/Token-Print. He also asks what's missing from the ecosystem for understanding and debugging LLMs.
More from Research
- ICLR 2027 opens Google's Gemini-powered Paper Assistant Tool to submitters for free pre-review feedback — iclr_conf · 2026-09-11
- Adversarial examples and geometry: an ECCV 2026 paper — ducha_aiki · 2026-09-11
- CMU Researchers Unveil TAHI, a Human-Agent Interaction Framework for Expert-Grade AI Output — EchoShao8899 · 2026-09-11
- Robotic fly achieves straight, level flight and recovers from wind gusts — moyix · 2026-09-11
- Two Millennium Problems solved within a month? Viral visualization gets a clearer remake — lishali88 · 2026-09-10
- AWS introduces AEM, a turn-level metric to isolate cascading agent errors — AWS ML Blog · 2026-09-10