Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Language Modeling on WikiText-2 (test) (Context Length 4096)
Loading...
5.11
PPL (WikiText-2)
Baseline (Ours)
-1,194.6856
6,903.9347
15,002.555
23,101.1753
Feb 6, 2026
Feb 11, 2026
Feb 17, 2026
Feb 22, 2026
Feb 28, 2026
Mar 5, 2026
Mar 11, 2026
PPL (WikiText-2)
Updated 1mo ago
Evaluation Results
Method
Method
Links
PPL (WikiText-2)
Baseline (Ours)
Finetuned=No, LM Eval=...
2026.03
5.11
LLVQ [shape-gain, 2 bit gain]
Finetuned=Yes, LM Eval...
2026.03
5.48
LLVQ [spherical shaping]
Finetuned=Yes, LM Eval...
2026.03
5.6
PV-tuning
Finetuned=Yes, LM Eval...
2026.03
5.84
QTIP
Finetuned=Yes, LM Eval...
2026.03
5.86
Quip#
Finetuned=Yes, LM Eval...
2026.03
6.19
Gemma2-9B (Baseline)
MP=false, bits=16, Bac...
2026.02
6.36
ScaleBITS + RTN
MP=true, bits=3.1, Bac...
2026.02
6.74
LLVQ [shape-gain, 2 bit gain]
Finetuned=No, LM Eval=...
2026.03
6.83
AQLM
Finetuned=Yes, LM Eval...
2026.03
6.93
LLVQ [spherical shaping]
Finetuned=No, LM Eval=...
2026.03
7.61
RTN-g128
MP=false, bits=3.1, Ba...
2026.02
8.11
Quip#
Finetuned=No, LM Eval=...
2026.03
8.22
ScaleBITS + RTN
MP=true, bits=2.1, Bac...
2026.02
8.89
RTN-g128
MP=false, bits=2.1, Ba...
2026.02
30,000
Feedback
Search any
task
Search any
task