List your company

Informal-to-formal proving on miniF2F (val)

25.8Proven Theorems RateDeepSeekMath-Base
14.276817.268420.2623.2516Oct 16, 2023Nov 3, 2023Nov 22, 2023Dec 11, 2023Dec 29, 2023Jan 17, 2024Feb 5, 2024

Evaluation results

MethodLinks
DeepSeekMath-BaseSize=7B, Prompting=Few-shot, Prover=Isabelle + Sledgehammer2024.0225.8
LLEMMA-34bDecoding=greedy, Proof Assistant=Isabelle2023.1021.03
LlemmaSize=34B, Prompting=Few-shot, Prover=Isabelle + Sledgehammer2024.0221
LLEMMA-7bDecoding=greedy, Proof Assistant=Isabelle2023.1020.6
LlemmaSize=7B, Prompting=Few-shot, Prover=Isabelle + Sledgehammer2024.0220.6
MistralSize=7B, Prompting=Few-shot, Prover=Isabelle + Sledgehammer2024.0218.9
CodeLlamaSize=34B, Prompting=Few-shot, Prover=Isabelle + Sledgehammer2024.0218.5
Code Llama 34bDecoding=greedy, Proof Assistant=Isabelle2023.1018.45
Code Llama 7bDecoding=greedy, Proof Assistant=Isabelle2023.1016.31
CodeLlamaSize=7B, Prompting=Few-shot, Prover=Isabelle + Sledgehammer2024.0216.3
SledgehammerDecoding=greedy, Proof Assistant=Isabelle2023.1014.72