metadata
base_model: unsloth/Qwen3-4B-unsloth-bnb-4bit
pipeline_tag: text-generation
license: mit
tags:
- base_model:adapter:unsloth/Qwen3-4B-unsloth-bnb-4bit
- lora
- sft
- transformers
- trl
- unsloth
- citation-verification
- reference-checking
- academic-text-analysis
inference:
parameters:
temperature: 0.6
top_p: 0.95
max_new_tokens: 2048
ollama:
parameters:
num_ctx: 4096
temperature: 0.6
top_p: 0.95
SemanticCite-Checker-Qwen3-4B
A fine-tuned Qwen3-4B model specialized for citation verification and reference checking. This model analyses citation claims against reference text snippets and provides structured JSON output with classification, reasoning, and confidence scores.
Model Details
Model Description
This model is designed to verify the accuracy of citation claims by comparing them against reference text snippets. It classifies citations into four categories (SUPPORTED, PARTIALLY_SUPPORTED, UNSUPPORTED, UNCERTAIN) and provides detailed reasoning with confidence scores in structured JSON format.
- Developed by: Sebastian Haan
- Model type: Causal Language Model (Fine-tuned)
- Language(s) (NLP): English
- License: MIT
- Finetuned from model: unsloth/Qwen3-4B-unsloth-bnb-4bit
Uses
Direct Use
This model is intended for:
- Academic semantic citation verification
- Citation contectual understanding
- Automated reference validation in scholarly writing
Out-of-Scope Use
This model should not be used for:
- General text classification tasks outside of citation verification
- Legal document verification (requires domain-specific expertise)
- Medical or health-related claim verification
- Decision-making in high-stakes scenarios without human oversight