Q3.6-27B-GLM-5.1-DA

Q3.6-27B-GLM-5.1-DA (Qwen3.6 GLM Distilled-Abliterated) is a reasoning-focused model built on top of Qwen/Qwen3.6-27B through the prithivMLmods/Qwen3.6-27B-abliterated-rMAX base. The model is optimized for rich, detailed, and context-aware reasoning using GLM-5.1 distilled reasoning traces combined with advanced refusal direction analysis and ablation-based training strategies to reduce internal refusal behaviors while preserving strong reasoning and instruction-following performance.

This model is intended strictly for research and learning purposes. Due to reduced internal refusal mechanisms, it may generate sensitive or unrestricted content. Users assume full responsibility for how the model is used. The authors and hosting platform disclaim any liability for generated outputs.

Note: This model is experimental and may generate artifacts.

Key Highlights

• GLM-5.1 Distillation: Fine-tuned using distilled reasoning traces derived from GLM-5.1 reasoning generations for enhanced mathematical and logical reasoning capabilities.
• Distilled-Abliterated (DA): Applies refusal direction analysis and ablation-based strategies to reduce internal refusal behaviors while maintaining reasoning quality.
• Qwen3.6 Backbone: Built on top of Qwen/Qwen3.6-27B via prithivMLmods/Qwen3.6-27B-abliterated-rMAX for strong instruction-following and reasoning performance.
• Math-Focused Reasoning: Optimized using high-quality mathematical reasoning traces from curated GLM-5.1 datasets.
• Instruction + Reasoning Fusion: Handles instruction-following and complex multi-step reasoning tasks seamlessly.
• 27B Scale Performance: Delivers high-capacity reasoning suitable for advanced research and complex tasks.

Datasets Used and Training Details

Category	Details
Base Model	Qwen/Qwen3.6-27B
Intermediate Base	prithivMLmods/Qwen3.6-27B-abliterated-rMAX
Final Model Size	27B Parameters
Training Type	Distillation + abliteration
Objective	Preserve reasoning quality while reducing refusal behaviors and improving instruction-following reliability
Reasoning Dataset	Jackrong/GLM-5.1-Reasoning-1M-Cleaned (Subset-Math, 6000 random samples used)
Alignment / Evaluation Dataset	prithivMLmods/harm_bench
Training Pipeline	TRL (Transformer Reinforcement Learning)
Training Focus	Mathematical reasoning, structured thinking, long-chain reasoning, robustness across diverse prompts

Quick Start with Transformers

pip install transformers==5.8.0
# or latest
pip install git+https://github.com/huggingface/transformers.git

from transformers import Qwen3_5ForConditionalGeneration, AutoProcessor
import torch

model = Qwen3_5ForConditionalGeneration.from_pretrained(
    "prithivMLmods/Q3.6-27B-GLM-5.1-DA",
    torch_dtype="auto",
    device_map="auto"
)

processor = AutoProcessor.from_pretrained(
    "prithivMLmods/Q3.6-27B-GLM-5.1-DA"
)

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "text",
                "text": "Solve this math problem step-by-step: If a train travels 240 km in 3 hours, what is its average speed?"
            }
        ],
    }
]

text = processor.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

inputs = processor(
    text=[text],
    padding=True,
    return_tensors="pt"
).to("cuda")

generated_ids = model.generate(
    **inputs,
    max_new_tokens=512
)

generated_ids_trimmed = [
    out_ids[len(in_ids):]
    for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]

output_text = processor.batch_decode(
    generated_ids_trimmed,
    skip_special_tokens=True,
    clean_up_tokenization_spaces=False
)

print(output_text)

Intended Use

• Mathematical Reasoning Tasks: Deep multi-step math reasoning powered by GLM-5.1 distilled traces
• Instruction Following: Hybrid prompts requiring both instruction adherence and reasoning
• Red-Teaming & Alignment Research: Evaluating reduced-refusal systems and refusal direction analysis
• Local High-Performance Deployment: Multi-GPU or optimized inference setups
• Research on Abliteration: Studying the effects of ablation-based training on reasoning preservation

Limitations & Risks

Important Note: This model intentionally minimizes built-in safety refusals.

• Sensitive Content Risk: May produce unrestricted or controversial outputs
• User Responsibility: Requires careful and ethical usage
• High Compute Demand: Requires significant VRAM or optimized quantization for efficient inference
• Abliteration Trade-offs: Reduced refusal may impact safety alignment and output filtering