k-chirkunov/gemma4-e4b-claims-comparison
k-chirkunov/gemma4-e4b-claims-comparison is a 7.9 billion parameter Gemma-4-e4b-it fine-tuned model by k-chirkunov. It functions as a Dialectal Arabic Natural Language Inference (NLI) engine, designed to compare two Arabic claims and categorize their semantic relationship into one of seven predefined categories. The model utilizes a thinking channel for internal reasoning before outputting the final category, making it specialized for detailed semantic analysis of Arabic text.
Loading preview...
Overview
k-chirkunov/gemma4-e4b-claims-comparison is a fine-tuned version of google/gemma-4-e4b-it specifically engineered as a Dialectal Arabic Natural Language Inference (NLI) engine. This 7.9 billion parameter model is designed to compare two Arabic claims (claim_a and claim_b) and classify their semantic relationship into one of seven distinct categories. It employs an internal thinking channel to reason through the claims before providing a precise categorical output.
Key Capabilities
- Dialectal Arabic NLI: Specialized in understanding and comparing semantic relationships between Arabic claims.
- Seven Semantic Categories: Classifies relationships into
Equivalent,Contradictory,OverlapIncomparable,DifferentFacts,Uncertain,A-MoreSpecific, andB-MoreSpecific. - Reasoning Channel: Utilizes a
<|channel>thoughtmechanism for internal, English-based reasoning, enhancing transparency and accuracy. - Tie-Resolution Protocol: Resolves ambiguous classifications by strictly selecting the category with the lowest numerical index.
Good for
- Automated semantic comparison of Arabic text.
- Applications requiring fine-grained analysis of claim relationships in Dialectal Arabic.
- Research and development in Arabic NLP, particularly for NLI tasks.