Metraipolis › archive › dc_3fa11c216093

Off-Manifold Collapse in Guided Protein Language Models — arXiv:2608.18597

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.18597
License (as archived)
CC0-1.0
Archived
2026-10-18T13:53:18.936Z
SHA-256 of the archived text
3518631e599aa1609a34ba9acbf5a330220f9835ef0bed39e09ede3914030b17
Stable URL of this page
https://metraipolis.com/documento/dc_3fa11c216093

How to cite

Off-Manifold Collapse in Guided Protein Language Models — arXiv:2608.18597 — arXiv (en), https://arxiv.org/abs/2608.18597 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3fa11c216093 (Metraipolis archive dc_3fa11c216093, sha256 3518631e599aa1609a34ba9acbf5a330220f9835ef0bed39e09ede3914030b17)
BibTeX
@misc{dc3fa11c216093,
  title = {Off-Manifold Collapse in Guided Protein Language Models — arXiv:2608.18597},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.18597},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3fa11c216093; sha256 3518631e599aa1609a34ba9acbf5a330220f9835ef0bed39e09ede3914030b17},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.18597. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Protein language models are widely used priors for protein sequence design, and a growing body of work controls them at inference time as an alternative to fine-tuning. Such guidance faces a dilemma: mild enough to preserve natural activation statistics, it barely moves the property; strong enough to move it, the generations become progressively harder to fold. We show the failure has a specific and cheaply detectable signature, an off-manifold collapse of the model's own representations. Guided activations fall toward a region statistically indistinguishable from random amino-acid input, and the sequences degenerate to low complexity, yet the property oracle being optimized can still score these generations as a success. The optimized oracle can therefore fail to witness the collapse and, for solubility, can actively reward it, whereas structure and composition expose the failure. Because the failure is already visible in a finished candidate, we detect it at the output rather than modify the generator. We introduce a cheap density prior over natural protein activations and keep only the candidates that remain typical under it, a training-free post-hoc step we call Mahalanobis filtering. At matched guidance settings it improves both the property score and the structural plausibility of the sequences it keeps at negligible cost, without touching the generator, and transfers across different guidance methods. We release the activation statistic at this https URL

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-20 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.18597 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T13:50:58.918Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-20?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).