Nexesenex/Llama_3.1_8b_DeepDive_3_R1_Prev_v1.0

Overview

Nexesenex/Llama_3.1_8b_DeepDive_3_R1_Prev_v1.0 is an 8 billion parameter language model created by Nexesenex through a merge using the Model Stock method. It is built upon the NousResearch/DeepHermes-3-Llama-3-8B-Preview as its base model, integrating two additional models: Nexesenex/Llama_3.1_8b_Smarteaz_0.2_R1 and huihui-ai/Dolphin3.0-Llama3.1-8B-abliterated. This merging strategy aims to combine the strengths of its constituent models.

Key Capabilities

• Merged Architecture: Utilizes the Model Stock method to blend multiple pre-trained models, potentially leading to improved generalization and performance.
• Llama 3.1 Base: Benefits from the foundational capabilities of the Llama 3.1 architecture.
• General Language Tasks: Suitable for a broad range of applications requiring natural language understanding and generation.

Good For

• Experimentation with Merged Models: Developers interested in exploring the performance characteristics of models created via advanced merging techniques.
• General-Purpose LLM Applications: Use cases that require a capable 8B parameter model with a 32768 token context length.
• Benchmarking: Provides a merged model for comparative analysis against other Llama 3.1 derivatives and similar-sized models.

Performance Highlights

Evaluations on the Open LLM Leaderboard show an average score of 27.47. Specific scores include:

• IFEval (0-Shot): 71.01
• BBH (3-Shot): 30.73
• MMLU-PRO (5-shot): 27.12

Detailed evaluation results are available on the Open LLM Leaderboard.