loeclos/gpt-oss-sarcastic-finetuning-v2

TEXT GENERATIONPricing:Input $0.3 / Output $1.2Concurrent Unit Cost:1Model Size:20BQuant:FP8Context Size:32kPublished:Aug 19, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The loeclos/gpt-oss-sarcastic-finetuning-v2 is a 20 billion parameter GPT-OSS model developed by loeclos, fine-tuned for sarcastic responses. It was trained using Unsloth and Huggingface's TRL library, achieving 2x faster training speeds. This model is designed for applications requiring a large language model with a distinct sarcastic tone.

Loading preview...

Model Overview

The loeclos/gpt-oss-sarcastic-finetuning-v2 is a 20 billion parameter language model developed by loeclos. It is a fine-tuned version of the unsloth/gpt-oss-20b-unsloth-bnb-4bit base model, specifically optimized for generating sarcastic outputs. The model was trained with the assistance of Unsloth and Huggingface's TRL library, which contributed to a reported 2x acceleration in the training process.

Key Capabilities

  • Sarcastic Text Generation: Specialized in producing responses with a sarcastic tone.
  • Efficient Training: Leveraged Unsloth for faster fine-tuning.
  • Large Scale: Built upon a 20 billion parameter architecture, providing robust language understanding and generation.

Good For

  • Applications requiring a distinct, sarcastic voice.
  • Creative content generation where humor and irony are desired.
  • Exploring the nuances of fine-tuning large models for specific stylistic outputs.