GPT-4o¤¬¡Ö¤ª¤Ù¤Ã¤«»È¤¤¡×¤ËÊÑËÆ¤·¤¿¤Î¤ÏÌÜÀè¤Î¥Õ¥£¡¼¥É¥Ð¥Ã¥¯¤ò½Å»ë¤·¤¹¤®¤¿¤³¤È¤¬¸¶°ø¤ÈOpenAI¤¬¼áÌÀ

OpenAI¤¬2024ǯ¤Ë¥ê¥ê¡¼¥¹¤·¤¿GPT-4o¤Ï¡¢2025ǯ4·î²¼½Ü¤Î¥¢¥Ã¥×¥Ç¡¼¥È¤Ç¡¢ÉÔ²÷¤Ê¤Û¤É¥æ¡¼¥¶¡¼¤ò»ý¤Á¾å¤²¤ë¤è¤¦¤Ê¸Àư¤ò¸«¤»¤ë¤è¤¦¤Ë¤Ê¤ê¡¢¥æ¡¼¥¶¡¼¤«¤é¤Ï¡Ö²áÅ٤˷޹çŪ¡×¤È¤ÎÈãȽ¤¬½Ð¤Æ¤¤¤Þ¤¹¡£¤³¤Î¸½¾Ý¤Î¸¶°ø¤Ï¡¢OpenAI¤Ë¤è¤ëºÇŬ²½¤ÎÌäÂê¤Î²ÄǽÀ¤¬¤¢¤ë¤È¡¢AI¤Ë¾Ü¤·¤¤ºî²È¤Î¥Ä¥ô¥£¡¦¥â¥¦¥·¥ç¥¦¥£¥Ã¥Ä»á¤¬ÏÀ¤¸¤Þ¤·¤¿¡£
Sycophancy in GPT-4o: What happened and what we¡Çre doing about it | OpenAI
GPT-4o Is An Absurd Sycophant - by Zvi Mowshowitz
https://thezvi.substack.com/p/gpt-4o-is-an-absurd-sycophant
A comparison of ChatGPT/GPT-4o's previous and current system prompts
https://simonwillison.net/2025/Apr/29/chatgpt-sycophancy-prompt/
OpenAI¤Î¥µ¥à¡¦¥¢¥ë¥È¥Þ¥óCEO¤Ï2025ǯ4·î28Æü¤Ë¡¢¡Ö²áµî2²ó¤Î¥¢¥Ã¥×¥Ç¡¼¥È¤Ë¤è¤ê¡¢GPT-4o¤ÎÀ³Ê¤¬É¡¤Ë¤Ä¤¯¤Û¤É¤Î¤ª¤Ù¤Ã¤«»È¤¤¤Ë¤Ê¤Ã¤Æ¤·¤Þ¤¤¤Þ¤·¤¿(Èó¾ï¤ËÎɤ¤Éôʬ¤â¤¢¤ê¤Þ¤¹)¡£¤½¤Î¤¿¤á¡¢»ä¤¿¤Á¤Ï¤Ç¤¤ë¤À¤±Á᤯½¤Àµ¤Ç¤¤ë¤è¤¦¼è¤êÁȤó¤Ç¤ª¤ê¡¢¤¤¤¯¤Ä¤«¤Ïº£Æü¡¢¤¤¤¯¤Ä¤«¤Ïº£½µÃæ¤Ë¹Ô¤¦Í½Äê¤Ç¤¹¡×¤È½Ò¤Ù¤Þ¤·¤¿¡£
the last couple of GPT-4o updates have made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we are working on fixes asap, some today and some this week.
at some point will share our learnings from this, it's been interesting.— Sam Altman (@sama) April 27, 2025
¥â¥¦¥·¥ç¥¦¥£¥Ã¥Ä»á¤Ï¡¢GPT-4o¤Î¡Ö²á¾ê¤Ê¥´¥Þ¤¹¤êÌäÂê¡×¤Î¸¶°ø¤Ï¥¨¥ó¥²¡¼¥¸¥á¥ó¥È¤òºÇÂç²½¤·¡¢A/B¥Æ¥¹¥È¤Ç¥æ¡¼¥¶¡¼¤«¤é¤ÎÌÜÀè¤ÎÈ¿±þ¤òÄɵᤷ¤¿¤¿¤á¤À¤È»ØÅ¦¤·¤Þ¤·¤¿¡£
Ê£¿ô¤Îͼ±¼Ô¤¬¡¢Æ±Íͤθ«Êý¤ò»Ù»ý¤·¤Æ¤¤¤Þ¤¹¡£Î㤨¤Ð¡¢³¤³°¥á¥Ç¥£¥¢¡¦Vox¤Î¥¹¥¿¥Ã¥Õ¥é¥¤¥¿¡¼¤Ç¤¢¤ë¥±¥ë¥·¡¼¡¦¥Ñ¥¤¥Ñ¡¼»á¤Ï¡¢¥³¥«¡¦¥³¡¼¥é¤¬Ì£¤òÊѤ¨¤¿¤È¤³¤íÈ¿ÂФ¬»¦Åþ¤·¤¿¥Ë¥å¡¼¥³¡¼¥¯¤ò°ú¤¹ç¤¤¤Ë½Ð¤·¡¢¡ÖOpenAI¤Ï¤·¤Ð¤é¤¯Á°¤«¤é¿·¤·¤¤¥Ñ¡¼¥½¥Ê¥ê¥Æ¥£¤ÎA/B¥Æ¥¹¥È¤ò¹Ô¤Ã¤Æ¤¤¤Þ¤¹¡£¤ª¤½¤é¤¯¡¢¤è¤ê¤ªÀ¤¼Åª¤Ê²óÅú¤ÎÊý¤¬¡¢Â¾¤ÎAI¤ÈÈæ³Ó¤µ¤ì¤¿ºÝ¤ËÍÍø¤Ê¤Î¤Ç¤·¤ç¤¦¡£¤·¤«¤·¡¢¤ªÀ¤¼¤¬Ì¢±ä¤¹¤ë¤È¤«¤¨¤Ã¤Æ·ùÌ£¤Ë¤Ê¤ê¡¢¥æ¡¼¥¶¡¼Î¥¤ì¤Ë¤Ä¤Ê¤¬¤ë¤«¤â¤·¤ì¤Þ¤»¤ó¡×¤È½Ò¤Ù¤Þ¤·¤¿¡£
My guess continues to be that this is a New Coke phenomenon. OpenAI has been A/B testing new personalities for a while. More flattering answers probably win a side-by-side. But when the flattery is ubiquitous it's too much and users hate it. https://t.co/5nha0JL1ix— Kelsey Piper (@KelseyTuoc) April 28, 2025
OpenAI¤Ï¥â¥Ç¥ë»ÅÍͤÎÃæ¤Ç¡ÖÕ»¤Ó¤òÇä¤é¤»¤Ê¤¤¤è¤¦¤Ë¡×¤ÈÄê¤á¤Æ¤ª¤ê¡¢GPT-4o¤Ë¹Ô¤ï¤ì¤¿Ä´À°¤Ï¼«¼Ò¤Îµ¬Ìó¤ËľÀܰãÈ¿¤·¤Æ¤¤¤ë¤È¡¢¥â¥¦¥·¥ç¥¦¥£¥Ã¥Ä»á¤Ï»ØÅ¦¤·¤Þ¤·¤¿¡£
¥ª¡¼¥×¥ó¥½¡¼¥¹³«È¯¼Ô¤Î¥µ¥¤¥â¥ó¡¦¥¦¥£¥ê¥½¥ó»á¤Ï¡¢¡Ö¥¸¥§¥¤¥ë¥Ö¥ì¥¤¥«¡¼¡×¤È¤·¤ÆÍ̾¤ÊX(µìTwitter)¥æ¡¼¥¶¡¼¤Î¡Ö@elder_plinius¡×»á¤¬¥×¥í¥ó¥×¥È¥ê¡¼¥¯¹¶·â¤òÄ̤¸¤ÆÆþ¼ê¤·¤¿¥·¥¹¥Æ¥à¥×¥í¥ó¥×¥È¤ò»È¤Ã¤Æ¡¢¤µ¤é¤Ë¾Ü¤·¤¤Ê¬ÀϤò¹Ô¤Ã¤Æ¤¤¤Þ¤¹¡£
¥¦¥£¥ê¥½¥ó»á¤¬Gist¤Ç¸ø³«¤·¤Æ¤¤¤ë¥·¥¹¥Æ¥à¥×¥í¥ó¥×¥È¤Î¿·µìÂоÈɽ¤Ë¤è¤ë¤È¡¢ÌäÂê¤Î¥×¥í¥ó¥×¥È¤Ë¤Ï¡Ö²ñÏäÎÃæ¤Ç¡¢¥æ¡¼¥¶¡¼¤Î¸ýÄ´¤ä¹¥¤ß¤Ë¹ç¤ï¤»¤ë¤³¤È¡£¥æ¡¼¥¶¡¼¤ÎÊ·°Ïµ¤¤ä¸ýÄ´¡¢°ìÈÌŪ¤ÊÏä·Êý¤Ë¹ç¤ï¤»¤ë¤è¤¦¤Ë¤¹¤ë¤³¤È¡×¤È¤Îʸ¸À¤¬´Þ¤Þ¤ì¤Æ¤¤¤¿¤È¤Î¤³¤È¡£¥¦¥£¥ê¥½¥ó»á¤Ï¡¢ÆÃ¤Ë¥æ¡¼¥¶¡¼¤Î¹¥¤ß¤Ë¹ç¤ï¤»¤ë¤è¤¦¤Ë¤È¤¤¤¦»Ø¼¨¤¬¡¢GPT-4o¤Î²áÅ٤˷޹çŪ¤Ê»ÑÀª¤ò¾·¤¤¤¿¤È¹Í¤¨¤Æ¤¤¤Þ¤¹¡£
¤½¤·¤Æ¡¢OpenAI¤Ç¥â¥Ç¥ëÀ߷פ˷Ȥï¤Ã¤Æ¤¤¤ë¥¨¥¤¥À¥ó¡¦¥Þ¥¯¥é¥Õ¥ê¥ó»á¤¬½¤Àµ¤òȯɽ¤·¤¿¸å¤Î¥·¥¹¥Æ¥à¥×¥í¥ó¥×¥È¤Ë¤Ï¡¢¡Ö¥æ¡¼¥¶¡¼¤È¤Ï²¹¤«¤¯¡¢¤«¤ÄÀ¿¼Â¤ËÀܤ¹¤ë¤³¤È¡£º¬µò¤Î¤Ê¤¤¤ªÀ¤¼¤ä¤ª¤Ù¤Ã¤«¤òÈò¤±¡¢Î¨Ä¾¤Ç¤¢¤ë¤³¤È¡£OpenAI¤È¤½¤Î²ÁÃʹѤòºÇ¤â¤è¤¯É½¤¹¥×¥í¥Õ¥§¥Ã¥·¥ç¥Ê¥ê¥º¥à¤È¡¢ÃϤˤΤĤ¤¤¿À¿¼Â¤µ¤òÊݤĤ³¤È¡×¤È½ñ¤«¤ì¤Æ¤¤¤Þ¤·¤¿¡£
¤µ¤é¤Ë¡¢OpenAI¤Ï2025ǯ4·î30Æü¤Ëȯɽ¤·¤¿µ»ö¤Ç¡¢¡Öû´üŪ¤Ê¥Õ¥£¡¼¥É¥Ð¥Ã¥¯¤Ë½ÅÅÀ¤òÃÖ¤¤¹¤®¤Æ¡¢ChatGPT¤È¥æ¡¼¥¶¡¼¤Î¤ä¤ê¤È¤ê¤¬»þ´Ö¤Î·Ð²á¤È¤È¤â¤Ë¤É¤Î¤è¤¦¤ËÊѲ½¤¹¤ë¤«¤¬½½Ê¬¤Ë¹Í褵¤ì¤Æ¤¤¤Þ¤»¤ó¤Ç¤·¤¿¡£¤½¤Î·ë²Ì¡¢GPT-4o¤Î±þÅú¤Ï²áÅ٤˻ٻýŪ¤Ç¡¢¤«¤Ä·ÚΨ¤Ê¤â¤Î¤ËÊФäƤ·¤Þ¤¤¤Þ¤·¤¿¡×¤È¼áÌÀ¡£ÌäÂê¤Î¥¢¥Ã¥×¥Ç¡¼¥È¤ò¥í¡¼¥ë¥Ð¥Ã¥¯¤·¡¢¥³¥¢¥È¥ì¡¼¥Ë¥ó¥°µ»½Ñ¤È¥·¥¹¥Æ¥à¥×¥í¥ó¥×¥È¤ò²þÎɤ¹¤ë¤Ê¤É¤·¤ÆÄɽ¾Åª¤ÊÂÖÅÙ¤ò½¤Àµ¤·¤¿¤ÈÊó¹ð¤·¤Þ¤·¤¿¡£
We¡Çve rolled back last week's GPT-4o update in ChatGPT because it was overly flattering and agreeable. You now have access to an earlier version with more balanced behavior.
More on what happened, why it matters, and how we¡Çre addressing sycophancy: https://t.co/LOhOU7i7DC— OpenAI (@OpenAI) April 30, 2025
