Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Language Modeling on LAMBADA OpenAI (Accuracy, AR%)
Loading...
59.3
Accuracy
Teacher (Qwen3-4B)
1.164
16.257
31.35
46.443
Jun 2, 2026
Jun 3, 2026
Jun 4, 2026
Accuracy
AR (%)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
AR (%)
Teacher (Qwen3-4B)
Precision=BF16
2026.06
59.3
-
Muon + PC
Zero-shot=true, Model=...
2026.06
58.12
-
Muon baseline
Zero-shot=true, Model=...
2026.06
57.17
-
Recover-LoRA (OpenHermes-10k)
Recovery Training Data...
2026.06
55.6
93.3
AdamW + PC
Zero-shot=true, Model=...
2026.06
53.66
-
Recover-LoRA (Synthetic-10k)
Recovery Training Data...
2026.06
53.4
89.5
AdamW baseline
Zero-shot=true, Model=...
2026.06
50.26
-
Quantized Student (Qwen3-4B)
Quantization=W2-GateUp...
2026.06
3.4
-
Feedback
Search any
task
Search any
task