Embed Notice

HTML Code

<blockquote style="position: relative; padding-left: 55px;"><section><a href="https://mastouille.fr/users/okisayaka/statuses/114047369192880113">okisayaka (okisayaka@mastouille.fr)'s status on Sunday, 23-Feb-2025 07:41:08 JST</a><a href="https://mastouille.fr/@okisayaka" title="okisayaka@mastouille.fr"><img src="https://gnusocial.jp/avatar/78496-48-20221221143531.webp" width="48" height="48" alt="okisayaka" style="position: absolute; left: 0; top: 0;">okisayaka</a><div><a href="https://mastouille.fr/@okisayaka/114047354391486649" rel="in-reply-to">in reply to</a></div></section><article><p>SNSは極右の情報が勝つのに、LLMの推論はリベラルが優勢。これって何なんだろう？<br>ChatGPTの新モデルが少し「右傾化」したとの論文があるが、それでもPolitical compassではかなりリベラル〜左派に位置した結果になっているように見える。<br><a href="https://www.nature.com/articles/s41599-025-04465-z" rel="nofollow noreferrer">https://www.nature.com/articles/s41599-025-04465-z</a></p></article><footer><a rel="bookmark" href="https://gnusocial.jp/conversation/4614118#notice-9032948">In conversation</a><time datetime="2025-02-23T07:41:08+09:00" title="Sunday, 23-Feb-2025 07:41:08 JST">about 2 months ago</time> <span>from <span><a href="https://mastouille.fr/@okisayaka/114047369192880113" rel="external" title="Sent from mastouille.fr via ActivityPub">mastouille.fr</a></span></span><a href="https://mastouille.fr/@okisayaka/114047369192880113">permalink</a><h4>Attachments</h4><ol><li><article><header><div>Domain not in remote thumbnail source whitelist: media.springernature.com</div><h5><a href="https://www.nature.com/articles/s41599-025-04465-z?error=cookies_not_supported&amp;code=00e06567-363b-47f4-bf89-d3449fb67fec">“Turning right”? An experimental study on the political value shift in large language models - Humanities and Social Sciences Communications</a></h5><div> from <span>Gu, Chao</span></div></header><div>Humanities and Social Sciences Communications - “Turning right”? An experimental study on the political value shift in large language models</div><footer></footer></article></li></ol></footer></blockquote>

Corresponding Notice

Embed this notice
okisayaka (okisayaka@mastouille.fr)'s status on Sunday, 23-Feb-2025 07:41:08 JST okisayaka
in reply to
SNSは極右の情報が勝つのに、LLMの推論はリベラルが優勢。これって何なんだろう？
ChatGPTの新モデルが少し「右傾化」したとの論文があるが、それでもPolitical compassではかなりリベラル〜左派に位置した結果になっているように見える。
https://www.nature.com/articles/s41599-025-04465-z
In conversationabout 2 months ago from mastouille.frpermalink
Attachments
1. Domain not in remote thumbnail source whitelist: media.springernature.com
  “Turning right”? An experimental study on the political value shift in large language models - Humanities and Social Sciences Communications
  from Gu, Chao
  Humanities and Social Sciences Communications - “Turning right”? An experimental study on the political value shift in large language models

Public

Embed Notice

HTML Code

Corresponding Notice