List your company

Language Model Evaluation on Tulu-3 (dev)

60.85Knowledge ScoreEqual
22.463632.429342.39552.3607Jul 1, 2026

Evaluation results

MethodLinks
EqualBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0760.8564.5559.0353.6168.5853.4960.02
ODMBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0760.2565.6959.4244.2263.5954.8358
Qwen2.5-7B-InstructModel=Qwen2.5-7B-Instruct2026.0760.2452.3654.768.6744.3660.4756.8
RegMixBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0759.668.1263.3755.7658.0455.9460.14
CAUSALMIX-SBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0759.3562.8858.6364.4367.4760.9562.28
DMOBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0759.1563.760.6254.0570.2454.3560.35
GridBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0759.0868.7765.9961.5544.9256.4359.46
DoReMiBackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0758.0263.257.3557.2868.2152.7559.47
CAUSALMIX-ABackbone=Qwen2.5-7B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0757.1464.0358.5165.5268.2157.6561.84
Llama-3.1-Tulu-3-8B-SFTModel=Llama-3.1-Tulu-3-8B-SFT2026.0751.6367.3756.1457.9168.3946.3957.97
Qwen2.5-0.5B-InstructModel=Qwen2.5-0.5B-Instruct2026.0729.932.1435.8434.130.3111.1428.9
DoReMiBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0729.8131.722.933.435.325.3429.74
CAUSALMIX-SBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0729.4531.326.3732.438.8225.9530.71
GridBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0729.3431.8426.6835.0216.6417.2626.13
DMOBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0728.8630.9122.5130.6835.8626.1929.17
CAUSALMIX-ABackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0728.7330.6324.6730.4842.5143.4533.41
ODMBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0728.6530.7424.3534.6230.512.6126.91
ODMBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0728.5930.525.1836.0633.6410.2827.37
DoReMiBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0728.5831.1222.1433.6139.7426.5630.29
RegMixBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0728.5730.1927.0234.3224.2116.6526.82
GridBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0728.4631.9830.4133.520.1523.3827.98
EqualBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0728.3932.9823.9632.137.8927.7830.51
CAUSALMIX-ABackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0728.3731.3124.6729.5738.6326.9329.91
CAUSALMIX-SBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0728.3130.9627.6430.9742.5136.4732.81
EqualBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0728.0729.2821.5936.0646.9525.0931.78
DMOBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0728.0731.8822.4226.941.5941.3732.04
CAUSALMIX-ABackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0727.9329.6823.5627.7643.8150.9233.94
DMOBackbone=Qwen2.5-0.5B, Samples=400K, Mixture Source=tulu-3-sft-mixture2026.0727.7431.7524.2531.838.6341.6232.63
DoReMiBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0727.729.6620.5834.5342.8832.9331.38
EqualBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0727.6529.823.323333.8314.5727.03
ODMBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0727.530.5325.4133.126.0623.6227.7
RegMixBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0727.2930.0425.5331.2817.0132.5627.28
CAUSALMIX-SBackbone=Qwen2.5-0.5B, Samples=100K, Mixture Source=tulu-3-sft-mixture2026.0727.2730.2327.7734.9434.9414.8127.85
RegMixBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0724.929.9430.2525.4625.3222.5226.4
GridBackbone=Qwen2.5-0.5B, Samples=800K, Mixture Source=tulu-3-sft-mixture2026.0723.9424.429.6723.8117.9322.6431.17