Model Overview

Ogno-Monarch-Neurotic-7B-Dare-Ties is a 7 billion parameter language model developed by AtAndDev. It was created using the DARE TIES merging method, combining two distinct base models: bardsai/jaskier-7b-dpo-v5.6 and eren23/ogno-monarch-jaskier-merge-7b-OH-PREF-DPO. This merging strategy aims to synthesize the capabilities of its components, particularly those related to instruction following and preference alignment.

Key Characteristics

• Architecture: A merged model, combining two DPO (Direct Preference Optimization) fine-tuned models.
• Parameter Count: 7 billion parameters, offering a balance between performance and computational efficiency.
• Context Length: Supports a context window of 4096 tokens, allowing for processing moderately long inputs.
• Merging Method: Utilizes the dare_ties method from mergekit, with specific density and weight parameters applied to each base model.

Use Cases

This model is particularly well-suited for applications that benefit from:

• Instruction Following: Generating responses that adhere closely to given instructions.
• Preference Alignment: Producing outputs that are aligned with human preferences, inherited from its DPO-trained base models.
• General Text Generation: Capable of various language generation tasks where a 7B model is appropriate.