Mr. Chatterbox is a Victorian-era ethically trained model
by y1n0 on 3/31/2026, 2:26:15 AM
https://simonwillison.net/2026/Mar/30/mr-chatterbox/
Comments
by: fastball
[delayed]
3/31/2026, 6:48:18 AM
by: lovelearning
I thought the title meant the training data used was ethics content and ethical reasoning. Turns out "ethically trained" means the training data used doesn't violate copyright laws.
3/31/2026, 5:40:58 AM
by: heyethan
Looks like a model size issue, but the behavior already seems largely shaped by the data distribution.
3/31/2026, 6:35:35 AM
by: parpfish
after testing, i'm pretty sure that either a) i dont understand Victorian speech very well or b) a model with 340million parameters doesn't generate particularly coherent speech
3/31/2026, 5:12:44 AM
by: gen6acd60af
<p><pre><code> >Honestly, it’s pretty terrible. >But what a fun project!</code></pre>
3/31/2026, 6:42:01 AM
by: kgeist
Prior art: <a href="https://news.ycombinator.com/item?id=46590280">https://news.ycombinator.com/item?id=46590280</a><p>>TimeCapsuleLLM: LLM trained only on data from 1800-1875
3/31/2026, 5:50:18 AM