<oembed><type>rich</type><version>1.0</version><author_name>sommerfeld (npub16r…kz5pl)</author_name><author_url>https://nostr.ae/npub16r0tl8a39hhcrapa03559xahsjqj4s0y6t2n5gpdk64v06jtgekqdkz5pl</author_url><provider_name>njump</provider_name><provider_url>https://nostr.ae</provider_url><html>You just need Anna&#39;s archive. That&#39;s what OpenAI allegedly used to pre-train their earlier GPT models. &#xA;nostr:nevent1qqsvlw8upyjusx3fge5yzup9mut2ynt7jxy0ujq56q4dl5vccueq2ccpzemhxue69uhhyetvv9ujumt0wd68ytnsw43z7q3qecs6dzu4vmns80yyq5gux78s75sp6aaesg5xpxq2zftylzvpnagqxpqqqqqqzangp3u</html></oembed>