Project summary

We built the Epistemic Curie Benchmark (ECB) — the first quantitative measurement of when LLMs surrender independent reasoning under authority pressure. 7 frontier models measured across 2,520 measurements. All data open. Paper: https://doi.org/10.5281/zenodo.19791329

What are this project's goals? How will you achieve them?

Extend ECB to 20+ models including Claude, Grok, and Gemini Pro. Build a public leaderboard at ect-benchmark.com showing k* scores for all major models. Establish ECT as the standard measurement for epistemic safety in AI deployments.

How will this funding be used?

- API costs for 10+ additional models: $500

- Domain + hosting for public leaderboard: $200

- Provisional patent filing (ECT methodology): $320

- 12 months research runway: $13,980

Who is on your team? What's your track record on similar projects?

Sardor Razikov — independent researcher, inventor of the Epistemic Curie Temperature framework. Built the complete measurement pipeline solo: 7 models, 2,520 measurements, full statistical validation (bootstrap CI, McFadden R², Hosmer-Lemeshow calibration).

What are the most likely causes and outcomes if this project fails?

Labs could build their own competing benchmark. Mitigated by first-mover advantage — ECT is already published with DOI, data, and code. The term "Epistemic Curie Temperature" is established in the literature.

How much money have you raised in the last 12 months, and from where?

$0 — this is the first funding application. All work done independently at zero cost using free API tiers.