--- format: "aidr-story-markdown/v1" id: "7369fa607d8842884ef480530a3e08209102e50e36168c81d249f9f017a064de" canonical_url: "https://aidr.today/7369fa60?lang=en" title: "Ryan Schaeffer Says Goodfire's 2.9% Model Loss Represents 'Brutal' Drop in Training Compute" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-11-13T18:22:52.000Z" category: "Research" topics: ["llm"] source_urls: ["https://huggingnews.com/ai/ryan-schaeffer-says-goodfires-29percent-model-loss-represents-brutal-dro-43e75d14","https://marketbrief.now/ai/ryan-schaeffer-says-goodfires-29percent-model-loss-represents-brutal-dro-43e75d14"] summary: "AI researcher Ryan Schaeffer questioned Goodfire's conclusion regarding neural network interpretability. Schaeffer argued that a 2.9% increase in language model loss on an Alibaba Qwen3.6 model represents a massive lack of training efficiency. Goodfire found that removing the weakest 92.2% of neuron activations for each forward pass caused the slight loss increase." --- # Ryan Schaeffer Says Goodfire's 2\.9% Model Loss Represents 'Brutal' Drop in Training Compute > [Open the canonical story]() **Published:** 2026-11-13T18:22:52.000Z **Category:** Research **Topics:** llm ## Summary AI researcher Ryan Schaeffer questioned Goodfire's conclusion regarding neural network interpretability\. Schaeffer argued that a 2\.9% increase in language model loss on an Alibaba Qwen3\.6 model represents a massive lack of training efficiency\. Goodfire found that removing the weakest 92\.2% of neuron activations for each forward pass caused the slight loss increase\. ## Sources - [Story source]() - [Story source]()