prithivMLmods/Megatron-Corpus-14B-Exp

TEXT GENERATIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:14.8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Feb 4, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Megatron-Corpus-14B-Exp by prithivMLmods is a 14.8 billion parameter language model based on the Qwen 2.5 architecture, fine-tuned on a synthetic math corpus dataset. It specializes in enhancing chain-of-thought (CoT) reasoning, logical problem-solving, and structured data processing. The model supports a 128K token context length and excels in complex reasoning tasks, instruction-following, and multilingual text generation across 29+ languages.

Loading preview...

Megatron-Corpus-14B-Exp Overview

Megatron-Corpus-14B-Exp is a 14.8 billion parameter model built on the Qwen 2.5 architecture, developed by prithivMLmods. It is specifically fine-tuned on a synthetic math corpus to significantly improve its chain-of-thought (CoT) reasoning and logical problem-solving capabilities. This model is designed for advanced analytical tasks, offering robust performance in understanding complex contexts and generating structured outputs.

Key Capabilities

  • Advanced Reasoning & Logic: Optimized for multi-step problem-solving, logical deduction, and contextual analysis, particularly in mathematical and scientific domains.
  • Fine-Tuned Instruction Following: Capable of generating precise responses, structured outputs (e.g., JSON), and extended long-form text up to 8K tokens.
  • Long-Context Support: Handles input contexts up to 128K tokens, enabling processing of extensive documents and complex dialogues.
  • Multilingual Proficiency: Supports over 29 languages, including major global languages, for diverse international applications.
  • Code Generation & Debugging: Generates optimized code and assists in identifying programming errors.

Good For

  • Advanced Logical & Analytical Reasoning: Ideal for complex problem-solving, multi-step deductions, and cognitive reasoning tasks.
  • Mathematical & Scientific Computation: Suitable for theorem proving, complex calculations, and scientific knowledge retrieval.
  • Structured Data Analysis: Processes tables, JSON, and other structured formats for data-centric applications.
  • Extended Text Generation: Capable of producing research papers, instructional guides, and in-depth reports.
  • Multilingual Reasoning: High proficiency across 29+ languages for global applications requiring logical processing.

Finetunes
3 models