solvrays/solvrays-finetuned-pdf

TEXT GENERATIONConcurrency Cost:1Model Size:2.5BQuant:BF16Ctx Length:8kPublished:Apr 30, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

The solvrays/solvrays-finetuned-pdf model is a 2.5 billion parameter fine-tuning of google/gemma-2b-it by Solvrays, specifically engineered for Zero-Hallucination Technical Retrieval. It prioritizes factual documentation over generative speculation, trained on a proprietary dataset of technical and architectural documentation. With an 8192-token context length, this model excels at precise information extraction from documents, optimized for overlapping document segments and deterministic precision.

Loading preview...

Model Overview

This model, developed by Solvrays, is a specialized fine-tuning of google/gemma-2b-it with 2.5 billion parameters, designed for Zero-Hallucination Technical Retrieval. It has been trained on a proprietary dataset of technical and architectural documentation to ensure deep contextual grounding and factual accuracy, rather than speculative generation.

Key Capabilities

  • Technical Grounding: Prioritizes factual documentation, avoiding generative speculation.
  • Chunk-Aware Memory: Optimized for processing overlapping document segments using a 256-token window.
  • Deterministic Precision: Best utilized with do_sample=False for architectural accuracy in responses.
  • Hallucination Control: Trained to respond with 'Not Documented' or an empty response if information is not present in its internal weights, aiding verification.

Good For

  • High-precision information retrieval from technical and architectural documents.
  • Use cases requiring zero-hallucination and factual accuracy from specific documentation.
  • Applications where cross-verification with original PDF source material is critical, especially for numerical accuracy.

Limitations

  • Not intended for general conversation or creative writing tasks.
  • Requires a specific prompt construction to activate its 'Knowledge Retrieval' mode.