The open‑source llama.cpp library now supports classifier_activation settings for rerankers and changes default activation functions.
NoteBy OossaPublished by Oossa: 1 min read
The ggml‑org/llama.cpp project released a new version on 2026-10-08. The update adds support for the classifier_activation parameter used by reranker models. It also maps the classifier GELU activation to GELU_ERF, accepts tanh, and sets tanh as the default classifier activation, with ModernBERT falling back to GELU_ERF.
Developers can now fine‑tune reranker models more easily, potentially improving search relevance.
Models
The company’s new three‑agent Claude‑based reviewer found 73.43% of core‑claim errors, far above the previous best of 14.81%.
Models
Mistral's new trillion‑parameter model scores 38 on an independent index, but costs more per task than competing open models.
Models
OpenAI reported two influence campaigns that used AI‑generated content to plant false stories in real media and blocked the involved accounts.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
