Gemma-4-26B-A4B with 39-69% less thinking at same accuracy (bottlecapai.com via hn) gemma 3w
event
LLM Fine-Tuning Announcement
-
Oct 23, 2026 · 9 min read ThinkingCap: Gemma-4-26B-A4B — up to 69% and on average 39% less thinking, at unchanged accuracy TL;DR This is the fifth model in the ThinkingCap series. We reduced how much Google's gemma-4-26B-A4B-it thinks, wit…
-
We study the 305M-parameter audio encoder shared by Gemma 4 and EmbeddingGemma 2 (identical except for its final projection) for Saudi and Gulf Arabic speech recognition, without changing any of its weights. A frozen system trains only a 2…
-
Show HN: Reverse engineering the ANE to run Gemma 4 26B at 100 Tok/s (github.com via hn) gemma 3w
After checking out Turbo Fieldfare (a very interesting project that streamed experts from SSD on Apple Silicon Macs to run Gemma 4 26B with very little RAM usage), I tried to make this project much faster and succeeded! This project gets u…
-
Safeguard Flag on ML/Kaggle Workflow outside of claudes outputs. (www.reddit.comhttps) gemmafine-tuning 4w
I’m a machine learning engineer competing in the Gemma 4 fine-tuning challenge on Kaggle. I’m building a small synthetic software engineering dataset with a training/holdout split to fine-tune a small model.