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.

python

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.