Model Card for Averroes-R1

Model Details

Model Description

• Base Model: Qwen3-1.7B
• Language(s) (NLP): English
• License: MIT
• Task: Foundational Philosophical Reasoning with Chain-of-Thought (CoT)
• Model Type: Instruction-tuned model emphasizing logical and conceptual reasoning in philosophy
• Dataset: moremilk/CoT_Philosophical_Understanding

Uses

Direct Use

The model is designed for:

• Educational use in teaching and learning philosophy
• Supporting AI assistants and chatbots focused on structured reasoning and conceptual understanding
• Serving as a tool for structured philosophical explanation
• Enhancing automated reasoning systems in conceptual and abstract domains

It is not intended to replace human philosophical analysis or provide moral/personal advice.

Out of Scope

This model is not designed for:

• In-depth exploration of specialized or niche philosophical debates
• Providing personal philosophical advice or opinions
• Real-time discussion or analysis of ongoing philosophical issues
• Handling highly subjective or interpretive arguments lacking foundational grounding

Bias, Risks, and Limitations

• May simplify nuanced philosophical perspectives
• Not suitable for advanced research or subjective debate
• Outputs depend on prompt clarity; ambiguous inputs may yield incomplete reasoning
• Trained for foundational reasoning, not for exhaustive domain knowledge

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("khazarai/Averroes-R1")
model = AutoModelForCausalLM.from_pretrained(
    "khazarai/Averroes-R1",
    device_map={"": 0}
)

question = """
What is the existentialist dilemma of freedom, and how do concepts like responsibility, anguish, and bad faith relate to it, according to Sartre?
"""

messages = [
    {"role" : "user", "content" : question}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize = False,
    add_generation_prompt = True,
    enable_thinking = True,
)

from transformers import TextStreamer
_ = model.generate(
    **tokenizer(text, return_tensors = "pt").to("cuda"),
    max_new_tokens = 2200,
    temperature = 0.6,
    top_p = 0.95,
    top_k = 20,
    streamer = TextStreamer(tokenizer, skip_prompt = True),
)

Training Data

Scope

This model was fine-tuned on tasks emphasizing foundational philosophical reasoning, focusing on:

• Understanding key philosophical concepts across major branches (ethics, epistemology, metaphysics, logic, etc.)

• Explaining philosophical principles through clear examples and structured reasoning

• Highlighting the logical and conceptual steps behind philosophical inquiry

• Building a strong foundational understanding of philosophical thought

Illustrative Examples

• Explaining the difference between empiricism and rationalism

• Describing the reasoning behind the categorical imperative

• Analyzing simple logical fallacies within philosophical arguments

Emphasis on Chain-of-Thought (CoT)

The dataset explicitly teaches step-by-step reasoning, allowing the model to show intermediate thoughts when analyzing or explaining philosophical ideas.

Focus on Foundational Knowledge

Rather than diving into complex, specialized debates, the dataset helps build a broad, structured foundation for philosophical reasoning.