Content
@
https://warpcast.com/~/channel/aichannel
0 reply
0 recast
0 reaction
Dan Romero
@dwr.eth
what’s the equivalent of “[query] + site:reddit.com” for LLM prompts?
6 replies
2 recasts
24 reactions
Gabriel Ayuso
@gabrielayuso.eth
There are many hacks that can lead the model towards a weight space that has better answers depending on your needs. Role prompting is a common one. "Act as a principal security engineer". Asking for points of view from specific individuals or publications (hallucinations tend to happen more depend on source). Other reasoning related like breaking down by step, or "take a breath" and other such hacks. It's been almost a year since I was deep in this work though so things might've changed.
2 replies
0 recast
7 reactions
KMac🍌 ⏩
@kmacb.eth
https://exa.ai/use-case/rag
0 reply
0 recast
3 reactions
maurelian
@maurelian.eth
Just ask for it? I sometimes ask perplexity to tell me “according to redditors” or “what would people on Reddit recommend?”.
0 reply
0 recast
1 reaction
shazow
@shazow.eth
For a brief moment, it was my name. 🫠 https://warpcast.com/shazow.eth/0x761da418
0 reply
0 recast
0 reaction
Jason
@jachian
Use a more robust agentic scraper like AgentQL and you can probably do that query https://warpcast.com/jachian/0x42f03cc6
0 reply
0 recast
0 reaction
neon
@neonrover
it’s not well supported in tools unless you move to operator or deep research, most tools like perplexity rely on search results not site browsing. some sites enable one shot prompts such as https://x.com/sumanth_077/status/1875180575516242093?s=46 and some scraping services listed in the channel support headless browsing/full html scrapes
0 reply
0 recast
0 reaction
