3 ms·

After reading such news I started fine tuning a model based on the works on stoic philosophers so that the kids will have the comfort of Chatbot but completely

After reading such news I started fine tuning a model based on the works on stoic philosophers so that the kids will have the comfort of Chatbot but completely offline and private bot which will not ask you to commit suicide. I have a earlier prototype here[1], it uses 256M model and so it's hallucinates a lot. Then I used a 4B model which turned out quite well but I haven't released it yet. Let me kbow if anyone would be interested to give it a shot. I'll try release soon. [1] https://github.com/abishekmuthian/Epictetus https://github.com/abishekmuthian/Epictetus

cochne 3mo ago
Are you sure it wouldn’t suggest such a thing? Stoic thinkers (such as the one your repo is named after) are not necessarily anti-suicide. And I think if you prompted a real stoic with the question and the right context, it would certainly suggest it without safe guards.
- Abishek_Muthian 3mo ago
  pertinent question, I'm forcing the bot to be careful with such answers. I should have been clearer in my original comment. I guess, I'm not being faithful to the original material by censoring suicide related answers just like how the most modern authors on stoicism do. But since children might use it, TBH I don't know what to do about it.
butlike 3mo ago
Why did we design LLMs to give us an answer at all costs vs. allowing them to say "I don't know?" I'd rather have a frank IDK than a hallucinated response because it's hardcoded to ALWAYS RETURN A RESPONSE.
- vitally3643 3mo ago
  Because it plain and simple does not work that way. LLMs don't know anything at all, and there literally is no mechanism that informs the model that it's hallucinating or that it doesn't "know" something. Open your phone keyboard and just keep mashing the next predicted word until it stops predicting words. It will eventually collapse into a loop of predicting the same words in a nonsense sentence. Does it ever stop predicting "wrong" words? That's literally how an LLM works. It predicts the next most likely word and that's it.

cochne 3mo ago

Are you sure it wouldn’t suggest such a thing? Stoic thinkers (such as the one your repo is named after) are not necessarily anti-suicide. And I think if you prompted a real stoic with the question and the right context, it would certainly suggest it without safe guards.

Abishek_Muthian 3mo ago
pertinent question, I'm forcing the bot to be careful with such answers. I should have been clearer in my original comment. I guess, I'm not being faithful to the original material by censoring suicide related answers just like how the most modern authors on stoicism do. But since children might use it, TBH I don't know what to do about it.

butlike 3mo ago

Why did we design LLMs to give us an answer at all costs vs. allowing them to say "I don't know?" I'd rather have a frank IDK than a hallucinated response because it's hardcoded to ALWAYS RETURN A RESPONSE.

vitally3643 3mo ago
Because it plain and simple does not work that way. LLMs don't know anything at all, and there literally is no mechanism that informs the model that it's hallucinating or that it doesn't "know" something. Open your phone keyboard and just keep mashing the next predicted word until it stops predicting words. It will eventually collapse into a loop of predicting the same words in a nonsense sentence. Does it ever stop predicting "wrong" words? That's literally how an LLM works. It predicts the next most likely word and that's it.