Technical evidence note

HP2D 48-mer -23 folding result.

MEVENTA Creative Runtime v10 produced a valid coordinate artifact for the classic HP2D 48-mer benchmark sequence with energy -23.

The result was exported as JSON coordinates, rendered from the artifact, and independently revalidated.

-23validated energy
23non-consecutive H-H contacts
48sequence length
0validation violations
899cycle of final -23
HP2D 48-mer fold with energy -23 and 23 H-H contacts
Publication rendering of the validated HP2D 48-mer fold. The technical evidence package contains JSON, validation, run log, used source files and SHA-256 hashes.

Claim

What is claimed.

MEVENTA Creative Runtime v10 produced a valid coordinate artifact for the HP2D 48-mer sequence with energy -23, corresponding to 23 non-consecutive H-H contacts.

This matches the best-known reported energy level for this benchmark sequence.

Validation

  • Energy: -23
  • H-H contacts: 23
  • Valid fold: true
  • Violations: 0
  • Sequence length: 48
  • Coordinates: 48

Run reference

Candidate:

cand_1782371765506_5eolwv

Final -23:

cycle 899 / point 11274

Run setup

Local MacBook run. No benchmark-specific pretraining dataset or model checkpoint was used for this run.

Boundaries

What is not claimed.

  • This is not an AGI claim.
  • This is not a general protein-folding claim.
  • This is not a compute-normalized benchmark comparison.
  • This is not a claim that HP2D is equivalent to biological protein folding.
  • This is not a claim that MEVENTA outperforms every published method.

Evidence package

Open evidence package.

The package contains the original JSON artifact, SVG/PNG/PDF figures, validation JSON, run log, used source files and SHA-256 hashes.

Open GitHub evidence package

Open technical evidence figure as PDF

Download evidence ZIP

Literature context

Framed without overclaim.

Yang et al. report the same -23 energy level for the Istrail 48mer using an LSTM-DQN approach. The comparison here is energy-level parity on the same benchmark sequence, not a compute-normalized equivalence claim.

Liang and Wong's Evolutionary Monte Carlo work is a classical reference point for HP-lattice folding simulations.