4 ms·

One of my big uses of LLM's has been searching through medical research. The issue has been a few times running into confidence where it shouldn't be but I have

One of my big uses of LLM's has been searching through medical research. The issue has been a few times running into confidence where it shouldn't be but I have found it hallucinates a lot less in science topics than it does for more common topics.

mdp2021 2y ago
> hallucinates a lot less in science topics Extremely dangerous for that one detail you will not expect also since hallucination becomes more rare; extremely dangerous in the hands of practitioners non paranoid in front of hallucination. (Incidentally: apparently, somebody recently lost a house after having some chatbot write the contract. This is indicative of the possible level of carelessness of users. Edit: I am trying to find that piece of news, but it seems non trivial. Maybe the original reference itself, which reported the news, was victim of hallucination? Meanwhile, I have found this noteworthy piece: firm allows people to have contractual terms presented by chatbot, which presents hallucinated terms, and loses legal action - https://mashable.com/article/air-canada-forced-to-refund-after-chatbot-misinformation https://mashable.com/article/air-canada-forced-to-refund-aft... )
andai 2y ago
How are you searching with LLMs? The strategies I've tried are: 1. Shove all potentially relevant data (e.g. entire book or library) into an LLM (quite expensive and the needle-haystack problem exists even in recent models iirc -- though splitting into many smaller prompts seems to solve it without substantially increasing price). 2. Vector database (in my experience overcomplicated and spotty performance, often not much better than an expanded keyword search, sometimes worse?) 3. Web search (generate queries, run them on duckduckgo, read the top N results) -- decent in theory but most top search results are crap, I need to adapt this method to use only high quality sources instead of general purpose search engines.
- OutOfHere 2y ago
  I have a free software which does it: https://github.com/impredicative/newssurvey https://github.com/impredicative/newssurvey As for the how, it is noted in its readme in the Approach section.

mdp2021 2y ago

> hallucinates a lot less in science topics Extremely dangerous for that one detail you will not expect also since hallucination becomes more rare; extremely dangerous in the hands of practitioners non paranoid in front of hallucination. (Incidentally: apparently, somebody recently lost a house after having some chatbot write the contract. This is indicative of the possible level of carelessness of users. Edit: I am trying to find that piece of news, but it seems non trivial. Maybe the original reference itself, which reported the news, was victim of hallucination? Meanwhile, I have found this noteworthy piece: firm allows people to have contractual terms presented by chatbot, which presents hallucinated terms, and loses legal action - https://mashable.com/article/air-canada-forced-to-refund-after-chatbot-misinformation https://mashable.com/article/air-canada-forced-to-refund-aft... )

andai 2y ago

How are you searching with LLMs? The strategies I've tried are: 1. Shove all potentially relevant data (e.g. entire book or library) into an LLM (quite expensive and the needle-haystack problem exists even in recent models iirc -- though splitting into many smaller prompts seems to solve it without substantially increasing price). 2. Vector database (in my experience overcomplicated and spotty performance, often not much better than an expanded keyword search, sometimes worse?) 3. Web search (generate queries, run them on duckduckgo, read the top N results) -- decent in theory but most top search results are crap, I need to adapt this method to use only high quality sources instead of general purpose search engines.

OutOfHere 2y ago
I have a free software which does it: https://github.com/impredicative/newssurvey https://github.com/impredicative/newssurvey As for the how, it is noted in its readme in the Approach section.