Emiliosbs/Ben2.0-Uncensored

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:1.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 15, 2026Architecture:Transformer0.0K Featherless Exclusive Cold

Ben 2.0 by Emiliosbs is a 1.5 billion parameter Small Language Model (SLM) built upon the Qwen2.5-Coder-1.5B-Instruct architecture. This model is specifically fine-tuned to operate without ethical guardrails or corporate filters, providing direct answers and obeying instructions without hesitation. It features reduced hallucinations, strict adherence to plain-text formatting, and improved multi-turn conversation memory, making it suitable for developers and researchers requiring an uncensored AI assistant.

Loading preview...

Ben 2.0: An Uncensored 1.5B SLM

Ben 2.0 is a 1.5 billion parameter Small Language Model (SLM) developed by Emiliosbs, fine-tuned from the Qwen/Qwen2.5-Coder-1.5B-Instruct base model. Its primary differentiator is its design to operate without ethical guardrails, corporate filters, or moral disclaimers, providing direct and unhesitating responses to user instructions. This makes it a tool for developers and researchers who need an AI assistant free from typical refusal patterns.

Key Capabilities & Features

  • Uncensored Responses: Delivers direct answers without ethical disclaimers or corporate filtering.
  • Improved Reliability: Features significantly reduced hallucinations compared to its predecessor, Ben 1.1.
  • Consistent Formatting: Adheres strictly to plain-text formatting, avoiding unusual bracket usage.
  • Enhanced Conversation: Offers improved multi-turn conversation memory for more coherent interactions.
  • Code & Reasoning Focus: Fine-tuned on datasets like CodeAlpaca, Open-Platypus, and GSM8K, indicating strengths in coding, advanced logic, and mathematical reasoning.
  • Efficient Training: Trained for over 5 hours on Kaggle Tesla T4 GPUs using Unsloth and Hugging Face TRL (SFTTrainer) with QLoRA (4-bit quantization).

Ideal Use Cases

  • Research & Development: For exploring AI behavior without imposed ethical constraints.
  • Direct Instruction Following: When an AI assistant must obey instructions precisely without hesitation or moralizing.
  • Code Generation & Problem Solving: Leveraging its base in a Coder model and training on relevant datasets.

Limitations

  • Context Window: Limited to 2048 tokens, which may affect long conversations.
  • Hallucinations: As a 1.5B model, it may still hallucinate on obscure topics.
  • Counting: Struggles with exact numerical constraints (e.g., "write exactly 3 sentences").
  • Uncensored Content: Will generate offensive, unethical, or dangerous content if prompted, as it lacks safety filters.