liminal 🦠on Nostr: I connected text embeddings to kind 30041 for semantic search. Would be cool to see ...
I connected text embeddings to kind 30041 for semantic search. Would be cool to see not only summaries, but the specific events they were based on. If you sum up the distances to the events they come up, normalized, you'd be able to get guess of the percent influence each event had on the summary, compared to the rest of the other events.
Published at
2025-03-01 03:13:02Event JSON
{
"id": "410dd248f330b754744ec52d870f15d6c39aa676d67b8b29eb029ec21e695f75",
"pubkey": "dc4cd086cd7ce5b1832adf4fdd1211289880d2c7e295bcb0e684c01acee77c06",
"created_at": 1740798782,
"kind": 1,
"tags": [
[
"e",
"2276dd12dd459389ced906737c4395673a586c18b19bc1e8555454a1347eb09a",
"",
"root"
],
[
"e",
"2bc019cd32ec666412d9bb53f489aa2575a1fa448bb98ddc119fbe79c4c80a1f"
],
[
"e",
"aea20fea60c909f407c3440ade79eaa2bf9ad1c4c466efb86ebcb77dfaf48bcf",
"",
"reply"
],
[
"p",
"1bc70a0148b3f316da33fe3c89f23e3e71ac4ff998027ec712b905cd24f6a411"
]
],
"content": "I connected text embeddings to kind 30041 for semantic search. Would be cool to see not only summaries, but the specific events they were based on. If you sum up the distances to the events they come up, normalized, you'd be able to get guess of the percent influence each event had on the summary, compared to the rest of the other events.",
"sig": "8a03bdcd8b6f676b8e1e10f5dc391630832fbbe85d107c9eae1ace30d7738a596bbc113fc3bb8999d302586452ff420016f620910bdc23a0f5d14e35bcbaec26"
}
