NeuralKrishna-7B-slerp Overview

NeuralKrishna-7B-slerp is a 7 billion parameter language model developed by Kukedlc. It is a product of merging two distinct models, Kukedlc/Neural4gsm8k and Kukedlc/NeuralMaxime-7B-slerp, utilizing the slerp (spherical linear interpolation) merge method. This approach combines the strengths of its constituent models to enhance overall performance.

Key Capabilities

• Reasoning: Achieves 73.46% on the AI2 Reasoning Challenge (25-Shot).
• Common Sense: Scores 88.96% on HellaSwag (10-Shot) and 83.27% on Winogrande (5-Shot).
• General Knowledge: Demonstrates 64.62% on MMLU (5-Shot).
• Factuality: Attains 74.29% on TruthfulQA (0-shot).
• Mathematical Problem Solving: Performs well on GSM8k (5-shot) with a score of 70.13%.

Good For

• Applications requiring a balanced performance across various language understanding and generation tasks.
• Scenarios where a merged model approach is beneficial for combining specific model strengths.
• General-purpose text generation and conversational AI within its 4096-token context window.