List your company
ResearchFormerly WizWand

Language Modeling (EM) on Lambada

89.7EM AccuracyPaLM 540B
8.47629.56350.6571.737Apr 5, 2022Dec 5, 2022Aug 6, 2023Apr 6, 2024Dec 6, 2024Aug 7, 2025Apr 9, 2026
Updated 6mo ago

Evaluation results

MethodLinks
PaLM 540BEvaluation Protocol=Few-shot, Number of Shots=82022.0489.7
Megatron-Turing NLG 530BEvaluation Protocol=Few-shot, Number of Shots=152022.0487.2
PaLM 540BEvaluation Protocol=1-shot2022.0481.8
GLaM 62B/64EEvaluation Protocol=1-shot2022.0480.9
PaLMModel Size (Billions of Parameters)=540, Num Tokens (Billions of Tokens)=780, Training FLOP Count (Zettaflops)=2527.2, Evaluation Protocol=0-shot2022.0477.9
PaLM 540BEvaluation Protocol=0-shot2022.0477.9
ChinchillaEvaluation Protocol=0-shot2022.0477.7
ChinchillaModel Size (Billions of Parameters)=70, Num Tokens (Billions of Tokens)=1400, Training FLOP Count (Zettaflops)=588.0, Evaluation Protocol=0-shot2022.0477.4
PaLMModel Size (Billions of Parameters)=62, Num Tokens (Billions of Tokens)=795, Training FLOP Count (Zettaflops)=295.7, Evaluation Protocol=0-shot2022.0475.4
GopherModel Size (Billions of Parameters)=280, Num Tokens (Billions of Tokens)=300, Training FLOP Count (Zettaflops)=504.0, Evaluation Protocol=0-shot2022.0474.5
PaLMModel Size (Billions of Parameters)=8, Num Tokens (Billions of Tokens)=780, Training FLOP Count (Zettaflops)=37.4, Evaluation Protocol=0-shot2022.0469.5
DoGraphBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0414.5
RegMixBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0414
Dynamic Loss-BasedBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0413.4
DoReMiBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0412.7
Data Mixing LawBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0412
UniformBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0411.6
DOGEBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0411.6