MIST-Mini-8B-Thinking is the reasoning version of MIST-Mini-8B by olaverse. Trained with 4 phases of GRPO (Group Relative Policy Optimization) reinforcement learning to show its reasoning process before answering. MIST-Mini-8B (base): User: What is 15% of 280? Model: 42 MIST-Mini-8B-Thinking: User: What is 15% of 280? Model: 15% means 15/100 280 × 15 = 4200 4200 / 100 = 42 The answer is 42. rewardthinkformat: +0.5 for using tags rewardcorrectness: +1.0 for correct answer rewardreasoning_steps: +0.3 for structured steps — 🧠 Transparent Reasoning — shows thinking before answering — 📐 Strong Math — 95% accuracy on GSM8K after training — 🔍 Trustworthy — you can verify the reasoning — ⚡ Fast — 8B model, runs on consumer GPUs — 🔓 Unrestricted — follows all instructions — Temperature 0.6 (lower than base model) gives more consistent reasoning — and are plain
Модальности:
Генерация текста
Области применения:
Логика и рассуждение Диалог / чат
Задача: Генерация текста
Автор: olaverse
Теги: llama, reasoning, grpo, thinking, llama-3.1, mist, conversational, en
Лайков: 3 | Загрузок: 27
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
