PulseAugur
EN
LIVE 13:38:39

Anthropic used "destructive scanning" to train Claude AI

Anthropic employed a method called "destructive scanning" to gather high-quality training data for its AI model, Claude. This process involved acquiring large datasets, ideally predating 2022, to avoid the perceived negative impact of generative AI on current text. AI

IMPACT This data acquisition technique highlights the challenges and methods involved in training advanced AI models.

RANK_REASON The item describes a specific methodology used by an AI lab for data acquisition, which is a research-related activity. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — fosstodon.org →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Anthropic used "destructive scanning" to train Claude AI

COVERAGE [1]

  1. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    "Destructive scanning was Anthropic’s solution to a problem: in order to “train” Claude, Anthropic had to procure a large, high-quality language dataset, prefer

    "Destructive scanning was Anthropic’s solution to a problem: in order to “train” Claude, Anthropic had to procure a large, high-quality language dataset, preferably one created before 2022 and the corrupting influence of generative AI on contemporary text." # DestructiveScanning …