Get the latest tech news

Show HN: Steerling-8B, a language model that can explain any token it generates


We release Steerling-8B, an 8B-parameter causal diffusion language model that is interpretable by construction — its predictions are routed through concepts you can measure, audit, and control.

None

Get the Android app

Or read this on Hacker News

Read more on:

Photo of language model

language model

Photo of token

token

Photo of Steerling-8B

Steerling-8B

Related news:

News photo

Essence and accident in language model-assisted coding

News photo

One Token to rule them all – Obtaining Global Admin in every Entra ID tenant

News photo

AI's $344B 'Language Model' Bet Looks Fragile