Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Science Question Answering on ARC-E (test)
Loading...
79.5
Accuracy
FLAN-137B
31.556
44.003
56.45
68.897
Mar 5, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
FLAN-137B
setting=zero-shot
2026.03
79.5
Imagine-DeBERTa-v3-L
KB=Synthetic VQA+, set...
2026.03
79.1
Imagine-DeBERTa-v3-L (Retrieval)
KB=Synthetic VQA+, inf...
2026.03
78.9
Imagine-DeBERTa-v3-L
KB=Synthetic VQA, sett...
2026.03
76
CAR-DeBERTa-v3-L
KB=AbsAT, setting=zero...
2026.03
75.3
Z-LaVI (OPT-30B)
setting=zero-shot
2026.03
59.5
OPT-30B
setting=zero-shot
2026.03
58.2
Imagine-RoBERTa-L
KB=Synthetic VQA, sett...
2026.03
57.9
CAR-RoBERTa-L
KB=AbsAT, setting=zero...
2026.03
57
Z-LaVI (BART-L)
setting=zero-shot
2026.03
56.1
Imagine-GPT-2-L
KB=Synthetic VQA, sett...
2026.03
55.1
Z-LaVI (RoBERTa-L)
setting=zero-shot
2026.03
51.8
GPT-Neo-2.7B
setting=zero-shot
2026.03
49.6
GPT-J-6B
setting=zero-shot
2026.03
44.1
SMLM
KB=*, setting=zero-shot
2026.03
33.4
Feedback
Search any
task
Search any
task