# Multi-task Evaluation

StableLM Zephyr 3B is a 3-billion-parameter instruction-tuned model trained on public datasets, synthetic datasets, and Direct Preference Optimization (DPO), delivering excellent performance.

A lightweight sentence embedding model based on DistilBERT, suitable for various text processing tasks

Conan-embedding-v1 is a Chinese text embedding model developed by the Tencent BAC team, focusing on semantic representation and similarity calculation for Chinese text.

Stella 400M v5 is an English text embedding model that excels in multiple text classification and retrieval tasks.

An English text re-ranking model based on the ModernBERT pre-training architecture, developed by Alibaba's Tongyi Lab, supporting long-text processing up to 8192 tokens.

IBM Granite Embedding 30M English is a transformer-based English text embedding model developed and released by IBM.

GTE Multilingual Base is a multilingual sentence embedding model supporting over 50 languages, suitable for tasks like sentence similarity calculation.

This model has been evaluated on multiple tasks in the Chinese Text Embedding Benchmark (MTEB), including text similarity, classification, clustering, and retrieval tasks.

Venusaur is a sentence embedding model developed based on the Mihaiii/Bulbasaur foundation model, focusing on sentence similarity and feature extraction tasks.

Snowflake Arctic Embed L is a model focused on sentence similarity and feature extraction, suitable for various natural language processing tasks.

Snowflake Arctic Embed M is a sentence transformer model focused on sentence similarity tasks, capable of efficiently extracting text features and calculating similarity between sentences.

Korean instruction-optimized model based on EEVE-Korean-Instruct-10.8B-v1.0, trained using Direct Preference Optimization (DPO) method

A medium-scale Mixture of Experts (MoE) implementation based on Dolphin-2.6-Mistral-7B-DPO-Laser, with an average performance improvement of approximately 1 point in evaluations

tao-8k-origin is a model focused on sentence similarity calculation, supporting multiple similarity measurement methods and excelling in various Chinese evaluation datasets.

stella-large-zh-v2 is a Chinese model focused on sentence similarity calculation, supporting various semantic text similarity tasks and text classification tasks.

Piccolo is a Chinese base model specializing in various natural language processing tasks such as semantic text similarity (STS), classification, clustering, and retrieval.

A 7B-parameter LLaMa model fine-tuned on the Stanford Alpaca dataset, focusing on open-source instruction tuning

Polyglot-Ko-5.8B is a large-scale Korean autoregressive language model developed by EleutherAI's multilingual team, with 5.8 billion parameters trained on 863GB of Korean data.

A general-purpose Chinese sentence embedding model for calculating sentence similarity and semantic search tasks.

A Korean-pretrained BERT model developed by the KLUE benchmark team, supporting various Korean understanding tasks

A compact Korean RoBERTa model trained under the LASSL framework, suitable for various Korean natural language processing tasks.

This is a BERT base model for Romanian, case-sensitive, trained on a 15GB corpus.

A converted version of the 2.7 billion parameter dense model based on the paper 'Efficient Large-scale Language Modeling with Mixture of Experts' by Artetxe et al.