ConvexAI/Luminex-34B-v0.2

TEXT GENERATIONConcurrency Cost:2Model Size:34BQuant:FP8Ctx Length:32kPublished:Feb 17, 2024License:yi-licenseArchitecture:Transformer0.0K Cold

ConvexAI/Luminex-34B-v0.2 is a 34 billion parameter language model developed by ConvexAI, based on the SimpleSmaug-34b architecture with LaserRMT applied. This model is optimized for general reasoning and language understanding tasks, demonstrating strong performance across various benchmarks. With a 32,768 token context length, it is suitable for applications requiring extensive contextual comprehension and generation.

Loading preview...

Luminex-34B-v0.2 Overview

ConvexAI/Luminex-34B-v0.2 is a 34 billion parameter language model developed by ConvexAI. It is built upon the SimpleSmaug-34b architecture and incorporates LaserRMT technology, enhancing its capabilities for complex language tasks. The model features a substantial context length of 32,768 tokens, allowing it to process and generate longer, more coherent texts.

Key Capabilities & Performance

Luminex-34B-v0.2 demonstrates strong general reasoning and language understanding, as evidenced by its performance on the Open LLM Leaderboard. Key evaluation results include:

  • Average Score: 77.19
  • AI2 Reasoning Challenge (25-Shot): 74.49
  • HellaSwag (10-Shot): 86.76
  • MMLU (5-Shot): 76.55
  • TruthfulQA (0-shot): 70.21
  • Winogrande (5-shot): 83.27
  • GSM8k (5-shot): 71.87

These scores indicate its proficiency in diverse areas such as common sense reasoning, multiple-choice question answering, and mathematical problem-solving.

Ideal Use Cases

This model is well-suited for applications requiring robust language understanding and generation, particularly those benefiting from a large context window. Its strong benchmark performance suggests it can be effectively used for:

  • Advanced text generation and summarization
  • Complex question answering and information extraction
  • Reasoning-intensive tasks and logical problem-solving
  • Applications where understanding long-form content is crucial