Compare Tülu 3 vs. Zochi

Tülu 3 is a cutting-edge language model created by the Allen Institute for AI (Ai2) that aims to improve proficiency in fields like knowledge, reasoning, mathematics, coding, and safety. It is based on the Llama 3 Base and undergoes a detailed four-stage post-training regimen: careful prompt curation and synthesis, supervised fine-tuning on a wide array of prompts and completions, preference tuning utilizing both off- and on-policy data, and a unique reinforcement learning strategy that enhances targeted skills through measurable rewards. Notably, this open-source model sets itself apart by ensuring complete transparency, offering access to its training data, code, and evaluation tools, thus bridging the performance divide between open and proprietary fine-tuning techniques. Performance assessments reveal that Tülu 3 surpasses other models with comparable sizes, like Llama 3.1-Instruct and Qwen2.5-Instruct, across an array of benchmarks, highlighting its effectiveness. The continuous development of Tülu 3 signifies the commitment to advancing AI capabilities while promoting an open and accessible approach to technology.

Zochi stands out as the first autonomous AI system capable of completing the entire scientific research cycle, ranging from formulating hypotheses to achieving peer-reviewed publication, while generating cutting-edge outcomes. In contrast to previous systems that were confined to specific, well-defined tasks, Zochi thrives in confronting research challenges that are at the cutting edge of artificial intelligence. The system's effectiveness is demonstrated through a series of peer-reviewed papers accepted at the ICLR 2025 workshops, highlighting Zochi's capacity to produce innovative and academically sound contributions. Furthermore, Zochi recognized a significant obstacle within the AI field: the issue of cross-skill interference during parameter-efficient fine-tuning. This problem arises when models are adapted for multiple tasks at once, leading to enhancements in one skill that may negatively impact others. To combat this challenge, Zochi introduced a novel approach called CS-ReFT (Compositional Subspace Representation Fine-tuning), which emphasizes the editing of representations instead of altering weights. This groundbreaking method has the potential to revolutionize how AI systems are fine-tuned for diverse applications.