Ezzeri Esa savarin

Lean-Eval Benchmark: Cost Calculation Details

Setup

Benchmark: lean-eval (311 problems from Mathlib contributions)
Duration: 1 hour per model (timeout 3600)
Per-problem timeout: 300 seconds (5 minutes)
Ordering: solve_count_desc (easiest problems first)
Effort: low (all models)
Attempts: 1 per problem

Simplified revisions for three prove2me documents

1. Cutoff-90 mission description

Original: 102 lines, 6 sections. Section 4 ("The next research question") belongs in the below-90 invitation, not here. Section 3 ("A route through the supplied proof") has useful content but mixes operational detail with the mathematical route.

Revised cutoff90.md

Formalize the supplied proof at cutoff 90

Claims and evidence: dimension-independent Hlawka constants for Schatten norms

This file pairs every factual sentence in our vibemathed entry with the exact place that backs it. Every repository link points to one fixed commit of savarin/hlawka-schatten, 1829591, so the line numbers cannot drift.

What is claimed, and what is not. Audenaert and Kittaneh's Problem 7 asks for the best dimension-independent Hlawka constant, if one exists. This repository settles two things:

Existence, for every p. A constant exists exactly when 1 < p < ∞.
The best constant for diagonal matrices, when p ≥ 256.

The best constant for general matrices remains open.

Context for reviewer

What this is

A proposed GitHub comment on leanprover/comparator#58 ("Comparator generates different instance names for problem and solution files"), cross-linking #59 ("Proof irrelevance is not respected in statements").

The reviewer who assessed our previous draft recommended this approach over filing a new issue.

Why a comment on #58, not a new issue

Steering-to-Eval Pipeline: Full Session Summary

Session: 2026-07-01 04:40–08:21 UTC Resume: claude --resume 6b8dec2f-08d2-4996-9624-b06eaee2071d Slack thread: #reporting/steering-experiment

What we did

Built and validated a complete pipeline for converting ephemeral human corrections to Cy into persistent, measured improvements. Scanned 9 Slack channels, found 15 steering threads, triaged all 15, ran 8-round steering conversations with Cy, wrote eval cases, ran baseline + override comparisons, and graduated 2 SOUL rules with eval-backed evidence.

Cy Behavior Steering Audit — June 24–July 1, 2026

What this is

A systematic scan of every Slack channel our bot can read (9 channels), looking for threads where a human steers Cy toward better behavior through back-and-forth correction. "Steering" = human gives task → reviews output → provides corrective feedback → Cy produces improved version. One-shot requests, simple Q&A, and bug reports without resolution are excluded.

Inventory: 15 steering threads found

The gold-standard example (reporting channel)

Cy Eval Baseline — Split by SOUL Override

Date: 2026-06-30 Branch: ezzeri/eval-suite-canon-aligned (PR #18) Runtime: Hermes

PR #18 does two things: (1) ships a SOUL.md change (ghostwriting_voice_control rule), (2) builds the eval harness (criteria-count scoring, 52 evals). This scorecard separates runs WITH the SOUL override from runs WITHOUT it, so we can see the actual clean Hermes baseline.

	{
	"prepared": "2026-09-26",
	"status": "Analytic proof supplied; cutoff 90 not Lean-verified.",
	"source_files": [
	{
	"name": "eventual-scalar-bound.md",
	"sha256": "ac7b40ede703ddfbc180c76855067ef0f7fb64077e03f62a82ce5ae6be2352a5"
	},
	{
	"name": "report-b8.md",

	{
	"prepared": "2026-09-26",
	"status": "Analytic proof supplied; cutoff 90 not Lean-verified.",
	"source_files": [
	{
	"name": "eventual-scalar-bound.md",
	"sha256": "ac7b40ede703ddfbc180c76855067ef0f7fb64077e03f62a82ce5ae6be2352a5"
	},
	{
	"name": "report-b8.md",

	{
	"prepared": "2026-09-26",
	"status": "Analytic proof supplied; cutoff 90 not Lean-verified.",
	"source_files": [
	{
	"name": "eventual-scalar-bound.md",
	"sha256": "ac7b40ede703ddfbc180c76855067ef0f7fb64077e03f62a82ce5ae6be2352a5"
	},
	{
	"name": "report-b8.md",