p-e-w/Qwen3-4B-Instruct-2507-heretic-v3-quantized-processing

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Feb 14, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Warm

The p-e-w/Qwen3-4B-Instruct-2507-heretic-v3-quantized-processing model is a 4 billion parameter instruction-tuned causal language model, based on the Qwen3-4B-Instruct-2507 architecture developed by Qwen. This version is a decensored variant created using Heretic v1.1.0, specifically optimized to reduce refusals while maintaining strong performance across general capabilities like instruction following, logical reasoning, and coding. It features a substantial 262,144 token native context length and excels in subjective and open-ended tasks, making it suitable for applications requiring less restrictive content generation.

Loading preview...

Overview

This model, p-e-w/Qwen3-4B-Instruct-2507-heretic-v3-quantized-processing, is a decensored version of the original Qwen3-4B-Instruct-2507, created using the Heretic v1.1.0 tool. It is a 4 billion parameter instruction-tuned causal language model from the Qwen family, designed to operate in a "non-thinking mode" without generating <think></think> blocks. A key differentiator of this Heretic-processed version is its significantly reduced refusal rate, dropping from 99/100 in the original to 9/100, while maintaining strong performance across various benchmarks.

Key Capabilities

  • Decensored Output: Significantly reduced content refusals compared to the base model.
  • Enhanced General Abilities: Demonstrates improvements in instruction following, logical reasoning, text comprehension, mathematics, science, coding, and tool usage.
  • Extensive Context Length: Supports a native context length of 262,144 tokens, enabling processing of very long inputs.
  • Multilingual Support: Shows substantial gains in long-tail knowledge coverage across multiple languages.
  • User Alignment: Markedly better alignment with user preferences in subjective and open-ended tasks, leading to more helpful and higher-quality text generation.
  • Tool Calling: Excels in tool calling capabilities, recommended for use with Qwen-Agent.

Good for

  • Applications requiring less restrictive content generation and reduced refusals.
  • Tasks demanding strong instruction following and logical reasoning over very long contexts.
  • Multilingual applications and scenarios needing broad knowledge coverage.
  • Creative writing, open-ended text generation, and subjective tasks where user preference alignment is crucial.
  • Agentic workflows and tool-use scenarios, leveraging its robust tool-calling abilities.