List your company

Question Answering on HotpotQA distractor setting (test)

82.2Answer F1HGN-large
58.07264.33670.676.864May 16, 2019Jul 10, 2019Sep 3, 2019Oct 28, 2019Dec 22, 2019Feb 15, 2020Apr 10, 2020

Evaluation results

MethodLinks
HGN-largeSize=large2020.0482.2——88.5—74.2—
HGNContext encoding backbone=RoBERTa-large2019.1182.1969.2262.7688.4747.1174.21—
GSAN-largeSize=large2020.0481.6——88.7—73.9—
Longformer-largeSize=large2020.0481.3——88.3—73.2—
Longformer2019.1181.256863.0988.3445.9173.16—
C2F Reader2019.1181.2467.9860.8187.6344.6772.73—
Quark (dev)Evaluation Split=dev2020.0481.2——87—72.3—
C2F Reader2020.0481.2——87.6—72.8—
ETC-large+Size=large2020.0481.2——89.1—73.6—
ETCScale=large2019.1181.1868.1263.2589.0946.473.62—
TAP 2 (ensemble)Ensemble=true2020.0479.8——86.7—70.7—
SAEScale=large2019.1179.6266.9261.5386.8645.3671.45—
SAE2020.0479.6——86.7—71.4—
EPS+BERT2019.1179.0565.7958.586.2642.4770.48—
TAP22019.1178.5964.9955.4785.5739.7769.12—
P-BERT2019.1174.1661.1851.3882.7635.4263.79—
ChainEx2019.1174.1161.2—————
LQR-Net2019.1173.7860.256.2184.0936.5663.68—
BERT Plusbackbone=BERT2019.0569.95642.380.626.958.1—
DFGNEntity Recognition Model=BERT NER2019.0569.6956.3151.581.6233.6259.82—
DFGN2019.1169.6956.3151.581.6233.6259.82—
DecompRC2019.1169.6355.2—————
DFGN + BERTbackbone=BERT2019.0568.555.249.981.131.958.2—
DFGNEntity Recognition Model=Stanford CoreNLP NER2019.0568.4955.1749.8581.0631.8758.23—
QFE2019.0568.153.957.884.534.659.6—
QFE*Status=unpublished2019.0568.0653.8657.7584.4934.6359.61—
QFE2019.1168.0653.8657.7584.4934.6359.61—
GRN*Status=unpublished2019.0566.7152.9252.3784.1131.7758.47—
GRN2019.0566.752.952.484.131.858.5—
KGNNSetting=distractor2019.1165.7550.8138.7476.7922.452.82—
Baseline ModelSource=Yang et al. (2018)2019.0559.0245.620.3264.4910.8340.16—
Yang et al. (2018)Setting=distractor2019.1159.0245.620.3264.4910.8340.16—
Baseline Model2019.1159.0245.620.3264.4910.8340.16—
Baseline2019.055945.620.364.510.840.2—
AGARBackbone=InternVL3.5-8B2026.06——————32.8
AGLABackbone=InternVL3.5-8B2026.06——————30.2
CGRBackbone=InternVL3.5-8B2026.06——————19.8
Plain VQABackbone=InternVL3.5-8B2026.06——————31.5
RandomBackbone=InternVL3.5-8B2026.06——————31.6
UniformBackbone=InternVL3.5-8B2026.06——————31.9
VARBackbone=InternVL3.5-8B2026.06——————25.9
VEA-BrBackbone=InternVL3.5-8B2026.06——————29.4
VEA-ConBackbone=InternVL3.5-8B2026.06——————29.6