ZTC Gate Tetris
🧱
35
One gate call: 2.1 s vs 0.0615 s. Watch it play out.
capturezone
capturezone
AI & ML interests
None yet
Recent Activity
liked a Space 5 minutes ago
FINAL-Bench/Tetris-JEV-LAYA-ZTC reacted to SeaWolf-AI's post with 🔥 5 minutes ago
The cost of a judging gate is usually quoted as a number. This puts it on a Tetris board.
Three boards get the same piece order, and on every move the same proposal and the same noise — a paired comparison. The gate decides one thing: keep this move, or draw again. Each board gets the same 60 seconds of gate time.
The text-writing gates get through 15–22 moves. The generation-free gate gets through 40–50. The boards that stop simply run out of clock.
It does not win on accuracy: on the same 2,018-question LODO set, JEV scores AUC 0.7350 against ZTC-Judge-27B's 0.7289. The separation is elsewhere. Clock — 2.1 s vs 0.0615 s per call, and on a 200-candidate agent screen one judging call measured 3.206 s generative vs 0.033 s readout, same server. Calibration — a gate is a threshold, and at ECE 0.4985 (vs ZTC 0.0245) a threshold stops carrying information. Mechanism — a text judge can name option 42 when there is no option 42; a scoring readout cannot. Not a lower error rate. No path.
The curve in the ZTC panel is real online fitting, scored prequentially — predict first, learn after — with base weights untouched. Not recursive self-improvement.
Limits, also stated on the page: Laya's AUC and latency are not our measurements and are set equal to JEV's, so calibration is the only measured axis it differs on. The page is a simulation driven by measured constants.
KO / EN / ZH.
https://huggingface.co/spaces/FINAL-Bench/Tetris-JEV-LAYA-ZTC
https://huggingface.co/FINAL-Bench/ZTC-Judge-27B liked a model 2 days ago
FINAL-Bench/Darwin-TTS-1.7B-CrossOrganizations
None yet
liked a
Space 5 minutes ago
reacted to SeaWolf-AI's post with 🔥 5 minutes ago
Post
The cost of a judging gate is usually quoted as a number. This puts it on a Tetris board.
Three boards get the same piece order, and on every move the same proposal and the same noise — a paired comparison. The gate decides one thing: keep this move, or draw again. Each board gets the same 60 seconds of gate time.
The text-writing gates get through 15–22 moves. The generation-free gate gets through 40–50. The boards that stop simply run out of clock.
It does not win on accuracy: on the same 2,018-question LODO set, JEV scores AUC 0.7350 against ZTC-Judge-27B's 0.7289. The separation is elsewhere. Clock — 2.1 s vs 0.0615 s per call, and on a 200-candidate agent screen one judging call measured 3.206 s generative vs 0.033 s readout, same server. Calibration — a gate is a threshold, and at ECE 0.4985 (vs ZTC 0.0245) a threshold stops carrying information. Mechanism — a text judge can name option 42 when there is no option 42; a scoring readout cannot. Not a lower error rate. No path.
The curve in the ZTC panel is real online fitting, scored prequentially — predict first, learn after — with base weights untouched. Not recursive self-improvement.
Limits, also stated on the page: Laya's AUC and latency are not our measurements and are set equal to JEV's, so calibration is the only measured axis it differs on. The page is a simulation driven by measured constants.
KO / EN / ZH.
FINAL-Bench/Tetris-JEV-LAYA-ZTC
FINAL-Bench/ZTC-Judge-27B
Three boards get the same piece order, and on every move the same proposal and the same noise — a paired comparison. The gate decides one thing: keep this move, or draw again. Each board gets the same 60 seconds of gate time.
The text-writing gates get through 15–22 moves. The generation-free gate gets through 40–50. The boards that stop simply run out of clock.
It does not win on accuracy: on the same 2,018-question LODO set, JEV scores AUC 0.7350 against ZTC-Judge-27B's 0.7289. The separation is elsewhere. Clock — 2.1 s vs 0.0615 s per call, and on a 200-candidate agent screen one judging call measured 3.206 s generative vs 0.033 s readout, same server. Calibration — a gate is a threshold, and at ECE 0.4985 (vs ZTC 0.0245) a threshold stops carrying information. Mechanism — a text judge can name option 42 when there is no option 42; a scoring readout cannot. Not a lower error rate. No path.
The curve in the ZTC panel is real online fitting, scored prequentially — predict first, learn after — with base weights untouched. Not recursive self-improvement.
Limits, also stated on the page: Laya's AUC and latency are not our measurements and are set equal to JEV's, so calibration is the only measured axis it differs on. The page is a simulation driven by measured constants.
KO / EN / ZH.
FINAL-Bench/Tetris-JEV-LAYA-ZTC
FINAL-Bench/ZTC-Judge-27B
liked 9
models 2 days ago
FINAL-Bench/Darwin-TTS-1.7B-Cross
Text-to-Speech • 2B • Updated • 111 • 53
FINAL-Bench/Darwin-4B-Opus
Text Generation • 8B • Updated • 75 • 30
FINAL-Bench/Darwin-9B-Opus
Text Generation • 10B • Updated • 88 • 43
FINAL-Bench/Darwin-35B-A3B-Opus-Q8-GGUF
Text Generation • 35B • Updated • 155 • 28
FINAL-Bench/Darwin-31B-Opus
Text Generation • 31B • Updated • 139 • • 65
FINAL-Bench/Darwin-35B-A3B-Opus
Image-Text-to-Text • 36B • Updated • 224 • 103
FINAL-Bench/Darwin-27B-Opus
Text Generation • 28B • Updated • 2 • 46
FINAL-Bench/Darwin-4B-David
Text Generation • 8B • Updated • 151 • 61
FINAL-Bench/Darwin-4B-Genesis
Text Generation • 8B • Updated • 244 • 67
liked 2
Spaces 2 days ago
liked a
dataset 2 days ago
liked a
Space 2 days ago
liked 4
models 2 days ago
upvoted a collection 2 days ago