Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Language Modeling Pre-training on C4 (val)

13.33PPL (60k)

FOAM-2

13.21813.97414.7315.486Jan 13, 2025Mar 8, 2025May 2, 2025Jun 26, 2025Aug 20, 2025Oct 14, 2025Dec 8, 2025
Updated 19d ago

Evaluation Results

MethodLinks
2025.12
13.33-15.7812.0111.3911.13270.2-------
2025.12
13.46-15.7112.2911.98-126.2-------
2025.01
14.88.54--13.21-279.10.532--19.7416.4113.7713.46
2025.12
15.17-18.4913.5412.812.63270.1-------
2025.12
15.4-18.1414.1314.02-297.3-------
2025.12
15.4-19.4914.0913.75-125.6-------
2025.01
15.49.28--13.75-274.40.541--20.7817.1914.413.9
2025.12
15.63-18.3514.3313.7713.69341.8-------
2025.12
15.69-18.6314.3614.19-130.1-------
2025.01
15.710.7--14.31-541.90.274--21.7917.4414.9914.52
2025.12
15.71-18.1613.8713.2313.19350-------
2025.12
15.98-18.4414.914.73-143.6-------
2025.01
15.989.28--14.73-335.40.526--20.4217.3615.1214.79
2025.12
16.13-18.7314.5613.413.23285.5-------
2025.06
--------2.96819.459----
2025.06
--------2.92318.596----
2025.06
--------2.93218.765----
2025.06
--------2.93418.803----
2025.06
--------2.92918.698----
2025.06
--------3.00320.145----
2025.06
--------2.9318.727----
2025.06
--------2.97619.609----
2025.06
--------2.91818.504----
2025.01
-16------------