[U] Update channel content
This commit is contained in:
+18
-1
@@ -2,11 +2,28 @@
|
||||
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="zh-cn">
|
||||
<id>https://aza.moe/life</id>
|
||||
<title>小桂桂的回忆录 📒</title>
|
||||
<updated>2025-04-28T08:22:02.546378+00:00</updated>
|
||||
<updated>2025-04-28T10:45:32.119242+00:00</updated>
|
||||
<link href="https://aza.moe/life" rel="alternate"/>
|
||||
<generator uri="https://lkiesow.github.io/python-feedgen" version="0.9.0">python-feedgen</generator>
|
||||
<logo>https://aza.moe/meru_256px.png</logo>
|
||||
<subtitle>「我们所经历的每个平凡的日常,也许就是连续发生的奇迹」</subtitle>
|
||||
<entry>
|
||||
<id>5287</id>
|
||||
<title>小桂桂的回忆录 📒 #5287</title>
|
||||
<updated>2025-04-28T10:23:37+00:00</updated>
|
||||
<content><p><span class="spoiler"><span>(是玩笑啦,实际上向量相近只是在语言应用中更相关而已,和“词意”并没有什么实际的关联... 比如「快」和「慢」即使是反义词,相似度却有 0.36 这么高,只是因为很有关联而已 qwq 不过还是挺好玩的hhh)</span></span></p></content>
|
||||
<link href="https://aza.moe/life?post=5287" rel="alternate"/>
|
||||
</entry>
|
||||
<entry>
|
||||
<id>5286</id>
|
||||
<title>小桂桂的回忆录 📒 #5286</title>
|
||||
<updated>2025-04-28T10:11:40+00:00</updated>
|
||||
<content><p>我让聊聊帮忙把房间里的碗洗了,然后聊聊看到了杯子没有洗,因为觉得杯子不是碗 <i class="custom-emoji" emoji-src="emoji/5445226814236533679.tgs">🧐</i></p>
|
||||
<p>...然后我们争论了好久杯子是不是碗</p>
|
||||
<p>为了一劳永逸证明杯子是碗,我搓了一个脚本,从 gemma 3 模型里掏出了向量化层,把两个词向量化之后计算余弦相似度。相似度数值越接近 1 词意就越相近</p>
|
||||
<p>然后计算了「杯子」和「碗」的相似度 = 0.29,「猫」和「动物」的相似度 = 0.16,杯子和碗的词意距离比猫和动物还要更近——如果杯子不是碗,猫就不是动物。但是猫肯定是动物,所以杯子也一定是碗 <i class="custom-emoji" emoji-src="emoji/5442995162114434162.tgs">☕</i> 证明结束</p></content>
|
||||
<link href="https://aza.moe/life?post=5286" rel="alternate"/>
|
||||
</entry>
|
||||
<entry>
|
||||
<id>5285</id>
|
||||
<title>小桂桂的回忆录 📒 #5285</title>
|
||||
|
||||
File diff suppressed because one or more lines are too long
Binary file not shown.
|
After Width: | Height: | Size: 82 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 19 KiB |
+160
-130
File diff suppressed because it is too large
Load Diff
+16
-1
@@ -12,7 +12,22 @@
|
||||
<link>https://aza.moe/life</link>
|
||||
</image>
|
||||
<language>zh-cn</language>
|
||||
<lastBuildDate>Mon, 28 Apr 2025 08:22:02 +0000</lastBuildDate>
|
||||
<lastBuildDate>Mon, 28 Apr 2025 10:45:32 +0000</lastBuildDate>
|
||||
<item>
|
||||
<title>小桂桂的回忆录 📒 #5287</title>
|
||||
<link>https://aza.moe/life?post=5287</link>
|
||||
<description><p><span class="spoiler"><span>(是玩笑啦,实际上向量相近只是在语言应用中更相关而已,和“词意”并没有什么实际的关联... 比如「快」和「慢」即使是反义词,相似度却有 0.36 这么高,只是因为很有关联而已 qwq 不过还是挺好玩的hhh)</span></span></p></description>
|
||||
<guid isPermaLink="false">5287</guid>
|
||||
</item>
|
||||
<item>
|
||||
<title>小桂桂的回忆录 📒 #5286</title>
|
||||
<link>https://aza.moe/life?post=5286</link>
|
||||
<description><p>我让聊聊帮忙把房间里的碗洗了,然后聊聊看到了杯子没有洗,因为觉得杯子不是碗 <i class="custom-emoji" emoji-src="emoji/5445226814236533679.tgs">🧐</i></p>
|
||||
<p>...然后我们争论了好久杯子是不是碗</p>
|
||||
<p>为了一劳永逸证明杯子是碗,我搓了一个脚本,从 gemma 3 模型里掏出了向量化层,把两个词向量化之后计算余弦相似度。相似度数值越接近 1 词意就越相近</p>
|
||||
<p>然后计算了「杯子」和「碗」的相似度 = 0.29,「猫」和「动物」的相似度 = 0.16,杯子和碗的词意距离比猫和动物还要更近——如果杯子不是碗,猫就不是动物。但是猫肯定是动物,所以杯子也一定是碗 <i class="custom-emoji" emoji-src="emoji/5442995162114434162.tgs">☕</i> 证明结束</p></description>
|
||||
<guid isPermaLink="false">5286</guid>
|
||||
</item>
|
||||
<item>
|
||||
<title>小桂桂的回忆录 📒 #5285</title>
|
||||
<link>https://aza.moe/life?post=5285</link>
|
||||
|
||||
Reference in New Issue
Block a user