Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-task Generalization on StoryCloze, OpenQA, ARC-E, ARC-C combined
Loading...
87.76
Average Accuracy
Trajectory
68.8216
73.7383
78.655
83.5717
Mar 1, 2026
Average Accuracy
ID Accuracy
OOD Average Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Average Accuracy
ID Accuracy
OOD Average Accuracy
Trajectory
Train Dataset=ARC-E, M...
2026.03
87.76
94.28
85.58
Trajectory
Train Dataset=ARC-C, M...
2026.03
84.05
85.24
83.66
Trajectory
Train Dataset=OpenQA,...
2026.03
84.04
88.2
82.65
Few-shot Accuracy
Model=Qwen2.5-14b
2026.03
77.2
-
-
Linear Probe
Train Dataset=ARC-E, M...
2026.03
75.14
84.32
72.08
Linear Probe
Train Dataset=ARC-C, M...
2026.03
74.96
75.6
74.74
Zero-shot Accuracy
Model=Qwen2.5-14b
2026.03
71
-
-
Linear Probe
Train Dataset=OpenQA,...
2026.03
69.55
80.31
65.97
Feedback
Search any
task
Search any
task