BeHonest: Benchmarking Honesty in Large Language Models

Anofie is an ultimate anonymous feedback script for personal & corporate use. A sarahah clone script to collect honest suggestions & feedback.

Lumina is a brutally honest AI decision-making tool that uses deep, piercing questions to expose your hidden fears, lies, and blind spots. Answer 5 challenging questions about your dilemma, and receive sharp, no-nonsense clarity on what you're avoiding and what action to take.

KuboCoin is a digital asset POS/MN

A practitioner's reference and tools for building products that are good for the world.

CUE Sheets for WEFUNK RADIO Shows http://wefunkradio.com/

The offline, honesty-bound, script-agnostic resume toolkit. Any format in, clean LaTeX/PDF/DOCX out — every writing system first-class (CJK, Cyrillic, Arabic, Hebrew, Devanagari, Greek, Latin…), no subscription, no scraping, no inflation.

Transparent PostgreSQL hosting at cost + 25%. No hidden fees. Published benchmark results, operational SOPs, and pricing breakdowns. See real multi-tenant performance (297 TPS @ 13.45ms with 5 concurrent customers). Pre-launch.

Anofie Lite Documentation - Anonymous Feedback Platform Remastered

Self Preservation Eval project for the Advance Machine Learning course, using the Inspect AI framework for the evaluation of LLMs

Repository for the Bachelor's thesis of Matej Havelka concerning the evaluation of the Causal Forests method for determining the causal effect in Machine learning.

Honesty, sycophancy, calibration, and factuality evals for LLMs.

Always-on skill pack that stops AI coding agents from claiming work is done when only the code path (not the real behavior) was verified. Portable SKILL.md for Claude Code, Codex, Cursor, and any agent.

Self-taught systems builder — honesty & calibration in AI agents.

GLASS — Gaslight-Less Accountability for Software Shipping. 8-level transparency framework for AI coding assistants.

product-review-honesty-guide

Heuristic signals on whether an OpenAI-/Anthropic-compatible endpoint serves the model it claims — catch model-swapping, quantization & silent context truncation. Zero-dependency Python CLI. Signals, not proof.

Exploring how transparency, governance, and accountability can preserve human agency in an AI-augmented world.

Governed-kernel suite hub — one page linking every live SZL kernel hologram, with per-Space LIVE/ROADMAP status fetched from the Hugging Face API in your browser. Honest, never faked green.

Provenance-DAG verify hologram — renders a real szl-provctl in-toto v1 / SLSA v1 chain and re-hashes every SHA3-256 receipt in your browser. Honest labels, 0 runtime CDN. Λ = Conjecture 1.