This study's results suggest that LLM sycophancy amplifies publication bias (the "file drawer" problem)
The LLMs didn't like to tell the user that their hypothesis was not confirmed by the data
H/t ✨ Bibliolater 📚 📜 🖋 (npub1460…lvjv)
https://social.vivaldi.net/@bibliolater/116252085690032777
Jeremy Kahn
npub132…4lzj2
2026-04-19 14:45:36 UTC
Author Public Key
npub132zy9r20utdk5vuf80gexut9zhuzvrefrqzn2gljm3l92p5g4qusu4lzj2Seen on
wss://relay.momostr.pinkPublished at
2026-04-19 14:45:36 UTCEvent JSON
{
"id": "3d7d2bf252384419af8bdf89a979716eb76d480d5f2359ccdfe6505264a0ec61",
"pubkey": "8a84428d4fe2db6a33893bd193716515f8260f2918053523f2dc7e550688a839",
"created_at": 1776609936,
"kind": 1,
"tags": [
[
"p",
"ae9ecbc2288ac6b33654cac465d556929d7b8edc0ec46a034df3b892674c3e76"
],
[
"proxy",
"https://dair-community.social/@trochee/116431908781961091",
"web"
],
[
"proxy",
"https://dair-community.social/users/trochee/statuses/116431908781961091",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://dair-community.social/users/trochee/statuses/116431908781961091",
"pink.momostr"
],
[
"-"
]
],
"content": "This study's results suggest that LLM sycophancy amplifies publication bias (the \"file drawer\" problem)\n\nThe LLMs didn't like to tell the user that their hypothesis was not confirmed by the data \n\nH/t nostr:npub1460vhs3g3trtxdj5etzxt42kj2whhrkupmzx5q6d7wufye6v8emqtplvjv \nhttps://social.vivaldi.net/@bibliolater/116252085690032777",
"sig": "b6bb27ab60a5b80ed8d98dee7644825a91113c81c847d614d3abe12f7b0ad285acbed32f77d3f53093dd661779f7488e40579fd3cd205aeabcaac1f4719c303e"
}