Skip to Main Content
Table 10: 

Intrinsic metrics results on EBMSASS 1000 and BIOSSES based on a model trained on STS-B. KL1 = KL-Divergence(pq), KL2 = KL-Divergence(qp): p = gold empirical distribution; q = predicted distribution.

EBMSASSBIOSSES
rCal / NLPDKL1 / KL2 ↓rCal / NLPDKL1 / KL2 ↓
LR MC 0.828 0.236 / 3.319 8.75 / 1.23 0.870 0.250 / 4.488 8.82 / 1.54 
ConvLR MC 0.806 0.201 / 4.668 12.74 / 1.46 0.823 0.304 / 12.59 19.64 / 2.06 
LR BBB 0.854 0.633 / 5351. 16297 / 5.00 0.836 0.530 / 11972 16598 / 4.90 
ConvLR BBB 0.806 0.736 / 1091. 2373.7 / 4.13 0.804 0.923 / 2076. 2631.2 / 5.01 
EBMSASSBIOSSES
rCal / NLPDKL1 / KL2 ↓rCal / NLPDKL1 / KL2 ↓
LR MC 0.828 0.236 / 3.319 8.75 / 1.23 0.870 0.250 / 4.488 8.82 / 1.54 
ConvLR MC 0.806 0.201 / 4.668 12.74 / 1.46 0.823 0.304 / 12.59 19.64 / 2.06 
LR BBB 0.854 0.633 / 5351. 16297 / 5.00 0.836 0.530 / 11972 16598 / 4.90 
ConvLR BBB 0.806 0.736 / 1091. 2373.7 / 4.13 0.804 0.923 / 2076. 2631.2 / 5.01 
Close Modal

or Create an Account

Close Modal
Close Modal