Artificial Intelligence · 20.08.2026, 04:31 UTC
Off-Manifold Collapse in Guided Protein Language Models
| Schweregrad | info |
|---|---|
| Kategorie | Artificial Intelligence |
| Quelle | arXiv cs.LG ↗ |
| Veröffentlicht | 20.08.2026 UTC |
Sicherheitsmeldung mit Schweregrad noch nicht bewertet. Technische Details im Tab „Originaltext“; empfohlene Schritte in der Checkliste.
arXiv:2608.18597v1 Announce Type: new Abstract: 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 https://huggingface.co/Shuibai12138/off-manifold-collapse-plm
Maßnahmen
⬇ Als MarkdownVerwandte Beiträge
- info Best GPU Neoclouds 2026: CoreWeave, Nebius, Lambda, Crusoe, and Groq Ranked by Published Pricing and Contracted Power
- info Anthropic brings Mythos 5 to its Claude Security vulnerability scanner
- info How agents can delegate better
- info Why API Test Generation Is a Judgment Problem, Not a Code Generation Problem