Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Zero-shot Evaluation on MMLU, ARC-C, PIQA, WinoG, GSM8K, HellaSwag, GPQA, and RACE

60.94Average Score

Base

29.157637.408845.6653.9112Feb 19, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
60.9468.2555.274.2767.7267.573.336.644.7
2026.02
54.5864.4644.9770.1864.0953.1568.3532.938.58
2026.02
54.464.1344.8870.1863.5452.8468.7232.4838.42
2026.02
51.6860.8442.8368.8259.7549.6665.1530.5235.85
2026.02
51.5860.941.6468.1260.3849.1365.430.8536.2
2026.02
31.8328.0221.2554.7351.931.5947.924.6524.55
2026.02
31.7226.9222.1854.5251.31.4448.224.324.9
2026.02
30.5224.121.0852.9952.251.1446.0523.7222.8
2026.02
30.3823.9620.6553.4352.171.0645.1223.523.15