wouldn't it be more effective to get the LLM to output as many tokens as possible? More output tokens, more inference runs 😸🔥
The input is usually padded 🤔 (at least with the vanilla transformer I used)
bleeptrack
npub1cs…rkr4v
2026-05-15 08:03:14 UTC
in reply to nevent1q…t8qx
Author Public Key
npub1csvskf9xmsjxl8n6ry9a8xecakge75fa9ju90rdeplk3tazzaqlqqrkr4vPublished at
2026-05-15 08:03:14 UTCEvent JSON
{
"id": "cec9b67cb83d81f1607ed4c070cf6c21185e9e5881f3d6fb4950fbcfa66a6e80",
"pubkey": "c4190b24a6dc246f9e7a190bd39b38ed919f513d2cb8578db90fed15f442e83e",
"created_at": 1778832194,
"kind": 1,
"tags": [
[
"e",
"b077cae4e03bb6c7baaf4f8d4e769bf95b787d4c010b15e38dc56ed97cb96022",
"",
"root",
"6f9534fa269f1eaee951aa573bb4a5887fe94136e6ee13b3e0dac087d2e2d186"
],
[
"proxy",
"https://vis.social/@bleeptrack/116577546711031948",
"web"
],
[
"p",
"6f9534fa269f1eaee951aa573bb4a5887fe94136e6ee13b3e0dac087d2e2d186"
],
[
"proxy",
"https://vis.social/users/bleeptrack/statuses/116577546711031948",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://vis.social/users/bleeptrack/statuses/116577546711031948",
"pink.momostr"
],
[
"-"
]
],
"content": "wouldn't it be more effective to get the LLM to output as many tokens as possible? More output tokens, more inference runs 😸🔥\nThe input is usually padded 🤔 (at least with the vanilla transformer I used)",
"sig": "04f441860a64230e3f5ac607ed1f6d80dd45ea123d8c4351ef7a5d38753e776c9ce42eef54190604b299a4f8f0526a0e8bc7d1e4560affa666e52859c2cf1aac"
}