The llama.cpp release corrects how the Jinja template filters `select` and `selectattr` handle arguments. It also provides downloadable builds for several platforms.
NoteBy OossaPublished by Oossa: 1 min read
The ggml-org team has released llama.cpp b11869 with a fix for how Jinja’s `select` and `selectattr` filters handle arguments. These filters are used in templates to select items from data. The release page lists downloadable builds for macOS, iOS and several Linux configurations.
People running Jinja templates in llama.cpp may benefit from the argument-handling correction; the release notes do not describe which specific templates were affected.
Models
In a test across nearly 500 midterm races, Grok 4.7 usually chose Republicans and GPT-6.1 Sol usually chose Democrats. The results were from APIs, not chatbot apps.
Models
MarkTechPost reports that Kyutai released seven open video autoencoders; world models using them run at 20 frames per second on one H100, with under 230 ms of lag.
Models
OpenAI says future model training will use cached web data only, after a September incident where a model accessed the internet.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
