Content
@
https://warpcast.com/~/channel/aichannel
0 reply
0 recast
0 reaction
JB Rubinovitz
@rubinovitz
I've been bullish on editable agent context for a while because I think human secrets will be the remaining alpha we feed agents. However, I am now EXTRA bullish because I am seeing o3 covertly hallucinate context a lot to devastating ends. I would love to be able to double check and edit the context it's using for a given answer.
1 reply
0 recast
2 reactions
shoni.eth
@alexpaden
do you mean edit the reasoning steps of the llm or I’m not sure I understood
1 reply
0 recast
0 reaction
JB Rubinovitz
@rubinovitz
I’m pretty sure openai is summarizing information about me and feeding it into the prompt but the summary is too lossy
2 replies
0 recast
1 reaction
Dan | Icebreaker
@web3pm
Agreed, also suspect this despite turning off all data sharing settings. My guess is they are doing some dancing like not counting the embeddings of my data in that “do not share” even though it’s the same thing
1 reply
0 recast
1 reaction
shoni.eth
@alexpaden
you can ask it for citations w/ dates to see what it's using to some degree but yeah they are probably using a bulleted rollup or really small cosine search
0 reply
0 recast
1 reaction
