# Multi-task Reasoning

Mi:dm 2.0 is an 'AI centered around South Korea' model developed using KT's proprietary technology, which deeply internalizes the unique values, cognitive frameworks, and common-sense reasoning of South Korean society.

GuardReasoner 1B is a version fine-tuned via R-SFT and HS-DPO based on meta-llama/Llama-3.2-1B, focusing on classification tasks for analyzing human-AI interactions.

A security protection model fine-tuned based on Llama-3.2-3B using R-SFT and HS-DPO methods, designed to analyze harmful content in human-computer interactions

GuardReasoner 8B is a fine-tuned model based on meta-llama/Llama-3.1-8B, specializing in reasoning-based LLM safety protection

A German-optimized large language model based on Meta Llama3-8B, continuously pre-trained on 65 billion German tokens, specifically optimized for German and supporting 32k long context

A large language model based on Google's Gemma 7B, fine-tuned using the Zephyr SFT recipe, primarily for text generation tasks

DareBeagle-7B is a 7B-parameter large language model obtained by merging mlabonne/NeuralBeagle14-7B and mlabonne/NeuralDaredevil-7B using LazyMergekit, demonstrating excellent performance across multiple benchmarks.

TinyLLaVA is a compact large-scale multimodal model framework focused on vision-language tasks, featuring small parameter size yet excellent performance.

A 6.7 billion parameter large language model based on the Galactica architecture, specializing in essay writing tasks, with an average score of 37.75 on the Open LLM Leaderboard.