Update README.md

README.md

@@ -3,9 +3,7 @@ license: mit
 ---
 ## Overview
 
-This model is a supervised fine-tuned (SFT) version of homebrewltd/Ichigo-llama3.1-s-base-v0.3, trained on over 1 billion tokens from the Instruction Speech WhisperVQ v4 dataset. This dataset builds upon Instruction Speech WhisperVQ v3, incorporating multi-turn speech conversations and advanced noise rejection capabilities.
-
-As a result, the Ichigo-llama3s models offer enhanced robustness to noisy environmental inputs and improved handling of multi-turn conversations, ensuring greater reliability and performance in real-world applications.
+Developed by Menlo Research, AlphaMaze is a novel model designed to enhance and assess visual reasoning in large language models (LLMs). Unlike approaches that rely on complex image generation, AlphaMaze uses a surprisingly simple task: solving text-based mazes. This requires the LLM to internally reconstruct the maze, plan its path, and strategically reset after dead ends. To further improve AlphaMaze's capabilities, we utilize the GRPO (Generalized Relative Policy Optimization) method. The AlphaMaze model itself offers a richer, more nuanced assessment of spatial understanding than traditional multiple-choice tests.
 ## Variants
 
 | No | Variant | Cortex CLI command |