Model Overview

nbeerbower/mistral-nemo-gutenberg-12B-v2 is a 12 billion parameter language model derived from the axolotl-ai-co/romulus-mistral-nemo-12b-simpo base model. It has been further finetuned using the jondurbin/gutenberg-dpo-v0.1 dataset, which implies a focus on improving instruction following and general text generation capabilities through Direct Preference Optimization (DPO).

Key Characteristics

• Base Model: Finetuned from axolotl-ai-co/romulus-mistral-nemo-12b-simpo.
• Training Method: Finetuned for one epoch using an A100 GPU on Google Colab, leveraging DPO techniques.
• Context Length: Supports a substantial context window of 32,768 tokens.

Performance Insights

Evaluations on the Open LLM Leaderboard indicate its performance across various benchmarks. While the average score is 24.05, it shows a notable IFEval (0-Shot) score of 62.03, suggesting proficiency in instruction following. Other scores include BBH (3-Shot) at 34.73 and MMLU-PRO (5-shot) at 27.77. These metrics provide a snapshot of its reasoning and general knowledge capabilities, positioning it as a model suitable for tasks requiring robust instruction adherence and text generation.