dubesor86

• Token impact by long-Chain-of-Thought Reasoning Models

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 21 comments

• LLM Chess tournament - Single-elimination (includes DeepSeek & Llama models)

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 12 comments

• Compared DeepSeek-R1 to DeepSeek-R1-Zero: surprising results

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 23 comments

• I ran o1-preview through my small-scale benchmark, and it scored nearly identical to Llama 3.1 405B

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 66 comments

• Dubesor LLM Benchmark table

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 7 comments

• LLama 3.1 405B Instruct - Top 5 Overall in my own testing

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 18 comments

• Small scale personal benchmark results (28 models tested)

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 31 comments

• Llama-3-70b: performs very similarly to Mistral Medium in my testing

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 7 comments

• Completely novel to me: im-also-a-good-gpt2-chatbot on LMSYS Arena using codeblock to draw Diagrams to supplement its explanations

  Posted by dubesor86@reddit | LocalLLaMA | View on Reddit | 10 comments