LLM Reference

LingoWhale 8B

About

LingoWhale-8B is a bilingual large language model constructed with a transformer architecture, featuring 8 billion parameters. Developed through collaboration between DeepLangAI and THUNLP Lab, this model was pretrained on an expansive dataset of high-quality Chinese and English texts. It excels in understanding and generating long-context text in these languages, providing services like question answering, content generation, bilingual translation, content summarization, and suggestions. Despite its strong abilities, LingoWhale-8B faces challenges like potential biases from its training data, weaker performance on complex mathematical calculations, and occasional hallucinations.

Capabilities

MultimodalFunction CallingTool UseJSON Mode

Specifications

Parameters8B
ArchitectureDecoder Only
Specializationgeneral