lean-laguna / results /humaneval_dflash.json
art87able's picture
Lean Laguna: Laguna XS.2 + DFlash — lossless single-GPU speedup + cheaper RL rollouts
8612587
raw
history blame contribute delete
688 Bytes
{
"kind": "greedy_byte_parity",
"compared": 14,
"mismatches": 0,
"lossless": true,
"decoding": "greedy (temperature=0)",
"method": "Each of 14 distinct mixed-difficulty prompts was completed by Laguna XS.2 with and without the DFlash speculator; the two outputs were compared byte-for-byte.",
"pass_at_1": null,
"pass_at_1_note": "HumanEval pass@1 was NOT run. Byte-level greedy parity is the strict superset guarantee (identical bytes => identical pass@1 by construction). A full HumanEval sweep is a documented next step.",
"also_lossless": "An earlier 20-prompt trivial run was also 0/20 lossless.",
"source": "HF Job 6a19d8b73a4b8cae6044dfdf (h200), 2026-05-29"
}