Sequence Scoring on Held-out text (G.4) benchmark leaderboard · SOTA2 Research

Sequence Scoring on Held-out text (G.4)

129.22Throughput (bytes/sec)pllama1B ◦ fα

Evaluation Results

Method	Links
pllama1B ◦ fα2026.03		129.22	0.9922	0.6877
pllama1B ◦ fα2026.03		90.74	0.9128	0.6327
pllama8B ◦ fα2026.03		88.65	0.8176	0.5667
pgpt2 ◦ fα2026.03		82.61	1.2434	0.8619
pllama8B ◦ fα2026.03		76.39	0.7392	0.5124
pllama8B ◦ fα2026.03		52.05	0.7153	0.4958
pgpt2 ◦ fα2026.03		49.1	1.1466	0.7948
pllama1B ◦ fα2026.03		47.81	0.8626	0.5979
pphi4 ◦ fα2026.03		42.39	0.8714	0.604
pphi4 ◦ fα2026.03		34.48	0.7748	0.5371
pllama8B ◦ fα2026.03		28.81	0.6959	0.4824
pgpt2 ◦ fα2026.03		28.46	1.0652	0.7383
pllama1B ◦ fα2026.03		26.15	0.8402	0.5823
pphi4 ◦ fα2026.03		24.89	0.7547	0.5231
pllama8B ◦ fα2026.03		16.86	0.6913	0.4791
pphi4 ◦ fα2026.03		16.55	0.7301	0.5061
pllama1B ◦ fα2026.03		14.89	0.8364	0.5797
pgpt2 ◦ fα2026.03		14.2	1.0385	0.7199
pphi4 ◦ fα2026.03		9.91	0.7241	0.5019
pllama8B ◦ fα2026.03		9.73	0.687	0.4762
pllama1B ◦ fα2026.03		6.8	0.8358	0.5793
pgpt2 ◦ fα2026.03		6.46	1.0293	0.7134
pphi4 ◦ fα2026.03		5.81	0.717	0.497
pllama8B ◦ fα2026.03		4.35	0.687	0.4762
pllama1B ◦ fα2026.03		2.62	0.8362	0.5796
pllama8B ◦ fα2026.03		2.6	0.6869	0.4761
pphi4 ◦ fα2026.03		2.55	0.7163	0.4965
pgpt2 ◦ fα2026.03		2.18	1.022	0.7084
pllama8B ◦ fα2026.03		1.57	0.6869	0.4761
pphi4 ◦ fα2026.03		1.54	0.7162	0.4965
pllama1B ◦ fα2026.03		1.42	0.8361	0.5796
pphi4 ◦ fα2026.03		0.86	0.7162	0.4964
pgpt2 ◦ fα2026.03		0.75	1.022	0.7084
pllama1B ◦ fα2026.03		0.73	0.836	0.5795
pgpt2 ◦ fα2026.03		0.35	1.02	0.707
pgpt2 ◦ fα2026.03		0.34	1.0025	0.6949