Why “rlcd” is being discussed
The most recent Hacker News stories and comments contributing to this topic's mentions.
> salty that he did not get recognition; claims RLCD without justification This salty meta is definitely written by a human.
by hbarka · Sep 22, 2026
If Jev is fundamentally trained using RLCD while you’re building on a Qwen model that was trained using RLHF, how can the resulting model be considered Jev-like?
by hbarka · Sep 21, 2026
Nobody outside TypeSafe even knows what RLCD is, so that’s impossible to answer.
by adastra22 · Sep 19, 2026
I read the nod to Typesafe Jev. Are you implying use of RLCD or still RLHF?
by hbarka · Sep 19, 2026
I don't really see the breakthrough in Jev. Classification, scoring, routing and returning probabilities over predefined choices are all established problems. We implemented cat…
by sreekanth850 · Sep 19, 2026
TypeSafe claims a new model architecture, a specialized "parallel sampler", and RLCD training specifically intended to make output probabilities calibrated. But no paper release…
by FooBarWidget · Sep 18, 2026
Interest
Proportion of Hacker News items mentioning "rlcd" over time.
Mentions
Total number of Hacker News items mentioning "rlcd" over time.