dataeaze/dataeaze-RegLLM-zephyr_7b_beta-dzcompli

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Jan 31, 2024License:cc-by-nc-sa-4.0Architecture:Transformer0.0K Open Weights Cold

The dataeaze/dataeaze-RegLLM-zephyr_7b_beta-dzcompli is a 7 billion parameter MistralForCausalLM-based model developed by dataeaze systems pvt ltd. It has been domain-adapted and instruction-finetuned specifically for regulatory compliance, with a current focus on Indian Banking rules and regulations. This model excels at providing precise answers to queries related to Indian banking regulations, making it suitable for RegTech applications. It is released under a non-commercial license for research purposes.

Loading preview...

Overview

The dataeaze/dataeaze-RegLLM-zephyr_7b_beta-dzcompli is a 7 billion parameter language model developed by dataeaze systems pvt ltd. It is built upon the MistralForCausalLM architecture, specifically finetuned from the zephyr-7b-beta model. The primary focus of this model is regulatory compliance, having undergone unsupervised pretraining and instruction finetuning for this domain. Currently, its expertise is concentrated on Indian Banking rules and regulations.

Key Capabilities

  • Specialized Regulatory Compliance: Provides precise and insightful answers to queries concerning Indian Banking regulations.
  • Domain-Adapted: Enhanced through unsupervised pretraining and instruction finetuning for the regulatory compliance domain.
  • Assistive AI: Designed to function as an assistive AI technology for compliance-related tasks.

Use Cases

  • Direct Use: Answering specific questions related to Indian Banking regulations.
  • RegTech Applications: Serving as a core component within broader RegTech solutions.

Limitations and Recommendations

The model is currently optimized for English language and Indian regulations. Its guardrails are inherited from the base Mistral/Zephyr models, which could be impacted by finetuning. Users are advised to use this model as an assistive tool and always verify information with source documents for decision-making. It is recommended to use the model with grounding on a set of regulatory documents for optimal performance and accuracy.

Finetunes
1 models