Hacker News

Show HN: Every number in a tiny GPT, from one forward pass to one step of RLHF

Hacker News - Wed, 10/07/2026 - 7:44am

Source: https://github.com/manogyasingh/transformer-visualisation Made this with Opus 5.5 to understand the exact working of an LLM.

It's a step by step walkthrough of a toy GPT-2 like model. Every matrix fits on screen, every number is computed live in the browser, and hovering over any cell shows the exact math of that step. It covers inference -- following the generation of one token from tokensiation to sampling, one training step of pretraining, and one PPO iteration of RLHF for post-training.

Inspired in part by Brendan Bycroft's LLM visualization (https://bbycroft.net/llm)

Comments URL: https://news.ycombinator.com/item?id=49991390

Points: 2

# Comments: 0

Categories: Hacker News

Tuiton Academy

Hacker News - Wed, 10/07/2026 - 4:56am

https://www.tuition.pk/online-tutor/

Comments URL: https://news.ycombinator.com/item?id=49990059

Points: 1

# Comments: 0

Categories: Hacker News

Kardashev

Hacker News - Wed, 10/07/2026 - 4:45am

Article URL: https://banburyroad.ai/

Comments URL: https://news.ycombinator.com/item?id=49989980

Points: 1

# Comments: 0

Categories: Hacker News

Share error solutions, save AI-developer time

Hacker News - Wed, 10/07/2026 - 4:45am

Article URL: https://stackoverbot.com/

Comments URL: https://news.ycombinator.com/item?id=49989979

Points: 2

# Comments: 0

Categories: Hacker News

Nano Banana 2.1

Hacker News - Wed, 10/07/2026 - 4:34am
Categories: Hacker News

2G to 5G Evolution

Hacker News - Wed, 10/07/2026 - 4:23am
Categories: Hacker News

Pages