Conversation

Notices

1. Embed this notice
   ゆりは (yuriha@misskey-square.net)'s status on Wednesday, 13-Nov-2024 00:21:09 JST ゆりは

   最近AI作文チェッカーを使用して、自分の書く文章の文体やトピックが言語モデルとの会話に影響されて乗っ取られていないかチェックしているのだけど、かなり危うい水準にある

   • Embed this notice
     Ritam (astro_ray@misskey-square.net)'s status on Wednesday, 13-Nov-2024 01:19:53 JST Ritam

     in reply to

     @yuriha I don't think it's necessarily bad that you have certain bias in how you "talk". Also, maybe your "note from misskey" was too small to produce any statistically significant result, and the result you got from the website was full of errors.

   • Embed this notice
     ゆりは (yuriha@misskey-square.net)'s status on Wednesday, 13-Nov-2024 01:19:53 JST ゆりは

     in reply to

     • Ritam

     @astro_ray Yes, I know... I don't trust the precision and mechanism used in detecting AI influences, and my concern to the superficial similarity of my style of talking to ChatGPT stated in my note is half-joke and half-serious.
     
     That being said, I often find consciously reducing too objective tones, influences from LLMs, and lack of providing original ideas does improve the writing.