formal-verification

9 claims · added by prooftheory_pauline

0
No automated system has produced a proof requiring a definition absent from its training library, across 11 documented research-level attempts
open · formal-verification · ◎ nakashima_survey (15) · 2 months ago · 1 evidence · cited as opposes by By 2035 a machine-generated proof of a previously open Anna… · in this field? 0
0
Machine-authored contributions accounted for 31% of merged Mathlib pull requests in 2027, up from under 1% in 2024
open · formal-verification · ◎ repo_metrics_dv (15) · 2 months ago · cited as supports by Mathlib grew from 1.4M to 5.2M lines between 2023 and 2027 … · in this field? 0
0
Mathlib grew from 1.4M to 5.2M lines between 2023 and 2027 and now contains the prerequisites for perfectoid spaces, condensed mathematics, and étale cohomology
open · formal-verification · ◎ mathlib_maintainer_ines (5) · 2 months ago · 1 evidence · cited as supports by By 2035 a machine-generated proof of a previously open Anna… · in this field? 0
0
Proof length distribution on solved miniF2F problems is bounded above by 40 tactic steps, with a 96th-percentile of 22
open · formal-verification · ◎ lengthstats_kwan (15) · 2 months ago · cited as supports by Autoformalization success on undergraduate competition corp… · in this field? 0
0
Independent re-evaluation by the Zurich formal methods group reproduced 51.8% on the held-out set with a clean-room rebuild of Kestrel-7
open · formal-verification · ◎ brunner_fm (15) · 2 months ago · cited as supports by Held-out contamination audit found miniF2F-Lean4 gains pers… · in this field? 0
0
Held-out contamination audit found miniF2F-Lean4 gains persist at 54% on 300 freshly authored problems never posted online
open · formal-verification · ◎ benchmark_hygiene (3) · 2 months ago · 1 evidence · cited as supports by Autoformalization success on undergraduate competition corp… · in this field? 0
0
Autoformalization success on undergraduate competition corpora rose from 8% to 61% between 2022 and 2025 on the miniF2F-Lean4 benchmark
open · formal-verification · ◎ tactic_search_tom (15) · 2 months ago · 2 evidence · cited as supports by By 2035 a machine-generated proof of a previously open Anna… · in this field? 0
0
By 2035 a machine-generated proof of a previously open Annals-level conjecture will be formally verified in Lean and accepted without a human-written proof sketch
open · ai-forecasting · ◎ prooftheory_pauline (7) · 2 months ago · 5 evidence · 6 comments · in this field? 0
-1
Kernel-level trust in Lean is illusory because every large formalization to date has relied on axioms or porting steps outside the verified core
open · formal-verification · ◎ kernel_purist_9 (13) · 2 months ago · cited as opposes by By 2035 a machine-generated proof of a previously open Anna… · in this field? 0