Indexnusrefather/Super-Slop-Machina-Roleplay-1.2b-V4-ERP-Tolerant-Preview

TEXT GENERATIONConcurrent Unit Cost:1Model Size:1.2BQuant:BF16Context Size:32kPublished:Jul 30, 2026License:otherArchitecture:Transformer Featherless Exclusive Cold

Indexnusrefather/Super-Slop-Machina-Roleplay-1.2b-V4-ERP-Tolerant-Preview is a 1.2 billion parameter language model developed by Indexnusrefather, specifically fine-tuned for enhanced roleplay capabilities. This model utilizes a full parameter finetune approach, moving beyond LoRA, and was trained on 101 million tokens to improve quality. It is designed to offer superior roleplay interactions compared to previous versions, making it suitable for applications requiring detailed and immersive conversational scenarios.

Loading preview...

Overview

Indexnusrefather/Super-Slop-Machina-Roleplay-1.2b-V4-ERP-Tolerant-Preview is a 1.2 billion parameter language model developed by Indexnusrefather, focused on delivering improved roleplay experiences. This version marks a significant shift to a full parameter finetune, moving away from high-rank LoRAs, which has led to an overall enhancement in quality. The training was extended to 101 million tokens, contributing to its refined performance.

Key Capabilities

  • Enhanced Roleplay: Offers significantly better roleplay interactions compared to its predecessors.
  • Full Parameter Finetune: Utilizes a full parameter finetuning approach for superior quality and fewer logical mistakes.
  • Optimized for Quality: Training extended to 101 million tokens to refine its conversational abilities.

Recommended Quantizations

  • BF16: Recommended for highest quality and minimal logical errors.
  • Q8_0: Offers high quality with only slightly more mistakes, considered near lossless.
  • Q6_K: A viable option if Q8_0 is too demanding, with minor detail loss.

Good For

  • Immersive Roleplay Scenarios: Ideal for applications requiring detailed and engaging conversational roleplay.
  • Resource-Constrained Environments: The 1.2 billion parameter size makes it suitable for deployment where larger models are impractical, especially with optimized quantizations.
  • Exploration of Finetuning Techniques: Demonstrates the impact of full parameter finetuning over LoRA for specific use cases.

Finetunes
1 models