d-michail/Mitos-Qwen3-1.7B-Dialogue

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 13, 2026License:cc-by-sa-4.0Architecture:Transformer Open Weights Featherless Exclusive Cold

d-michail/Mitos-Qwen3-1.7B-Dialogue is a 2 billion parameter Qwen3-based language model fine-tuned for Greek e-government question answering. It specializes in providing brief, clerk-like responses to inquiries about Greek administrative procedures, leveraging a dataset of 90,145 synthetic citizen-clerk dialogues. This model excels in domain-specific question answering, particularly when augmented with retrieval, and is intended for research in Greek public administration QA.

Loading preview...

Overview

d-michail/Mitos-Qwen3-1.7B-Dialogue is a 2 billion parameter model based on Qwen3-1.7B, specifically fine-tuned for question answering related to Greek e-government administrative procedures. It was trained on a unique dataset of 90,145 synthetic citizen-clerk dialogues, generated by Qwen3.8-27B from the National Registry of Administrative Procedures (Mitos).

Key Capabilities

  • Specialized Greek QA: Optimized to answer questions about Greek administrative procedures in a concise, clerk-like manner (25-45 words).
  • Domain-Specific Knowledge: Learns the register and format of Greek public administration, though factual accuracy is significantly enhanced with Retrieval-Augmented Generation (RAG).
  • Research Tool: Intended for research in closed-book and RAG-based Greek public administration question answering.

Good For

  • Developing Greek e-government chatbots: Provides a strong foundation for systems requiring precise, brief answers on administrative topics.
  • Researching RAG effectiveness: Ideal for experiments on how retrieval mechanisms improve factual recall in domain-specific, low-resource language contexts.
  • Understanding domain adaptation: Demonstrates fine-tuning a general-purpose LLM for a highly specific, structured knowledge domain.

Limitations

  • Not legal advice: Answers are not legally authoritative and require verification against live procedures.
  • Factual accuracy: Closed-book performance for factual recall is low; RAG is crucial for accuracy.
  • Synthetic data: Training data is synthetic, which may reproduce generator errors or phrasing biases.
  • Staleness: Based on a snapshot of the Mitos registry, so information can become outdated.

Finetunes
2 models