ChaoticNeutrals/Eris_Remix_DPO_7B

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Mar 7, 2024License:otherArchitecture:Transformer0.0K Cold

ChaoticNeutrals/Eris_Remix_DPO_7B is a 7 billion parameter language model developed by ChaoticNeutrals, fine-tuned with DPO for enhanced roleplay and chat capabilities. This model is designed to perform well in both Alpaca and ChatML formats, offering versatility for conversational AI tasks. It is optimized for generating engaging and coherent responses in interactive scenarios, building upon the Eris_Remix_7B base model.

Loading preview...

Model Overview

ChaoticNeutrals/Eris_Remix_DPO_7B is a 7 billion parameter language model developed by ChaoticNeutrals, a collaborative effort between @Jeiku and @Nitral. This model is a DPO (Direct Preference Optimization) remix of the base Eris_Remix_7B model, specifically fine-tuned to excel in conversational and roleplay-oriented tasks.

Key Capabilities

  • Optimized for Roleplay and Chat: The model has undergone DPO training for 200 steps over 1 epoch using a specialized dataset (athirdpath/DPO_Pairs-Roleplay-Alpaca-NSFW), enhancing its ability to generate engaging and contextually relevant responses in interactive scenarios.
  • Versatile Format Support: Eris_Remix_DPO_7B is designed to function effectively with both Alpaca and ChatML instruction formats, providing flexibility for integration into various applications.
  • Enhanced Conversational Coherence: The DPO fine-tuning aims to improve the model's ability to maintain coherent and natural dialogue flows, crucial for immersive roleplay and chat experiences.

Good For

  • Roleplaying Applications: Ideal for creating interactive story-driven experiences or character-based simulations.
  • General Chatbots: Suitable for developing engaging and responsive conversational agents.
  • Creative Writing Assistance: Can be leveraged for generating dialogue or narrative content within a conversational context.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p