Benchmarking LoRA Methods for Fine-Tuning LLMs on Financial Datasets
XBRL tagging is a key step in creating XBRL reports. Numerical entities in texts such as earning calls can to be tagged with US GAAP tags.
Usage
Input: Provide a sentence containing financial information.
Output: Key entities and their corresponding US GAAP (Generally Accepted Accounting Principles) tags will be generated by the base model and our fine-tuned model.
Llama 3.1 8b (base) output
Llama 3.1 8b (base) output
Llama 3.1 8b (fine-tuned for XBRL tagging) output
Llama 3.1 8b (fine-tuned for XBRL tagging) output
Entites
⋮
US GAAP tags
⋮
Entites
⋮
US GAAP tags
⋮
Ground Truth Answer
Ground Truth Answer
Entites
⋮
US GAAP tags
⋮
Entites
⋮
US GAAP tags
⋮
Examples
Pages:
1 2
Analyze an existing XBRL report with ease using our fine-tuned model as a chatbot. The model allows extraction of US GAAP tags, values, or financial formulas from the XBRL report.
Usage
Input: A financial question and an XBRL file name.
Output: The answer to the question will be generated by the base model and our fine-tuned model. Click on any numbers to locate the value in the XBRL report.