Anima-Core/an1-core
View on GitHub ↗ ★ 0
Core AN1 engine for meaning-field extraction and downstream task heads. Private repository for enterprise evaluation and pilot deployments.
5 mentions on Hacker News by 4 people: 0 stories, 5 comments · first 2025-12-10 · latest 2025-12-10
20082026
-
…. Your vibe-coded eval has cheated this to collapse it into a binary selection on row 46 in https://github.com/Anima-Core/an1-core/blob/main/experiments... , making the problem baseline 50% on random choice instead of 25%, making the…
comment by augment_me · 2025-12-10 · in Post-transformer inference: 224× compression of Llama-70B with improved accuracy -
…that: - The repo always runs the full teacher model to extract activations and uses them - see https://github.com/Anima-Core/an1-core/blob/main/an1_core/fi... - There are weird stub files, e.g. the Hellaswag repro doesn't actually have the…
comment by Tiberium · 2025-12-10 · in Post-transformer inference: 224× compression of Llama-70B with improved accuracy -
Here is a working link to the same paper: https://github.com/Anima-Core/an1-core/blob/main/papers/Post...
comment by bigtones · 2025-12-10 · in Post-transformer inference: 224× compression of Llama-70B with improved accuracy -
…but with an empirical implementation. I just open-sourced the reference version here: GitHub: https://github.com/Anima-Core/an1-core Paper + DOI: https://zenodo.org/records/17873275 It isn’t about bigger models. It’s about reorganizing the…
comment by anima-core · 2025-12-10 · in The Solution Will Come from the Field -
…paper, statistical results, and methodology. A reference implementation (non-optimized) is here: https://github.com/Anima-Core/an1-core Production variants (AN1-Turbo, FPU work, etc.) are not included. I’m an outsider to academia so I’m…
comment by anima-core · 2025-12-10 · in Post-transformer inference: 224× compression of Llama-70B with improved accuracy