Incompressible Knowledge Probes: Estimating Black-Box LLM Parameter Counts via Factual Capacity
Bojie Li · 2026
Read Incompressible Knowledge Probes: Estimating Black-Box LLM Parameter Counts via Factual Capacity by Bojie Li (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-24827.txt.
Introduction
Li introduces Incompressible Knowledge Probes, a benchmark of 1,400 factual questions. The questions span 7 tiers of obscurity. They isolate knowledge that resists derivation by reasoning and compression by better architecture. Closed-source labs do not disclose parameter counts. Yet storing facts requires weights, so recall lower-bounds parameter count. Scoring with no hallucination penalty removes a hyperparameter and a flooring choice. Calibrating a log-linear mapping on 93 open models, the instrument recovers capacity to within a factor and rankings. Its 90 percent interval spans about 3 times in either direction.
Catalog record
| CREATOR | Bojie Li |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 191024 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.24827