Feature the LFM Space, evolved Qwen model, and Ternary Bonsai kernels

README.md

@@ -17,19 +17,9 @@ We study how learned systems represent, adapt, and generalize, with experiments
 
 ## Explore our work
 
-### MONARCH · Language models on your GPU
-
-How far can careful GPU programming take a small language model in a browser?
-
-MONARCH explores WebGPU and WGSL kernel optimization for **Liquid AI's LFM2.5-230M**. The public demo runs inference on your device, with editable prompts, Markdown responses, live tokens per second, and a repeatable benchmark. The release includes kernel source, build instructions, model checksums, and documented measurement conditions.
-
-**[Try the demo](https://huggingface.co/spaces/inductiveML/monarch-webgpu)** · [Read the experiment](https://inductive.ml/experiments/monarch) · [Browse the code](https://huggingface.co/spaces/inductiveML/monarch-webgpu/tree/main) · [Model files](https://huggingface.co/inductiveML/LFM2.5-230M-MONARCH)
-
-### Staged Learned Coordinates · Teaching trees to understand geometry
-
-Can a small neural network learn a coordinate system that makes a decision tree's job easier? This experiment studies learned representations in front of XGBoost to simplify curved and rotated decision boundaries.
-
-[Read the experiment](https://inductive.ml/staged-learned-coordinates)
+- **[LFM2.5 WebGPU Space · MONARCH](https://huggingface.co/spaces/inductiveML/monarch-webgpu)** — Run LFM2.5-230M in your browser with our WGSL kernels, generate text locally, and measure your GPU's tokens per second.
+- **[Qwen3.6-35B-A3B · Evolved Mixed-Bit](https://huggingface.co/inductiveML/Qwen3.6-35B-A3B-evolved-mxbit)** — A 12.63 GB MLX quantization with per-module precision selected by evolutionary search. Runs on Apple Silicon with stock `mlx-lm`.
+- **[Ternary Bonsai kernels · Ternel](https://github.com/inductiveML/ternel)** — Custom Metal kernels that execute Ternary Bonsai 27B directly from losslessly packed 1.75-bit weights on Apple Silicon. [Get the MLX model](https://huggingface.co/inductiveML/Ternary-Bonsai-27B-mlx-lossless-1.75bpw).
 
 ## How we work