AI personality labels varied with prompts and model compression
Preprint: ENFJ was the leading MBTI label across tested open-source models, while reported types varied with prompts, precision and decoding settings.
Search
2,230 results
1401–1420
AI models often get the science right but miss the rules
An arXiv preprint benchmark of 2,527 scientific questions finds that correctness and instruction following often diverge.
Misaligned nanowire shells steer plasmonic hot spots
Preprint: An arXiv study finds that nonconcentric shells give access to otherwise dark resonances and move near-field hot spots toward thinner regions.
New theorem maps curvature, flow and topology in hyperbolic space
Preprint: Under a stated curvature condition, closed hypersurfaces in hyperbolic space admit a terminating flow with surgery and a restricted topological classification.
A guide to elliptic-curve ranks leaves key puzzle open
Preprint offers a concise introduction to two ways of defining rank for elliptic curves, while stressing that the central conjecture is still unproved.
REBELS-25 shows warm and cold gas sharing large-scale rotation
Preprint: ALMA and JWST observations of REBELS-25 find shared rotation across warm and cold gas, while 55% to 98% of star formation is obscured in four regions.
Preprint tests method for separating exposure effects from time trends
Simulations found the design stayed unbiased under one pattern of time-induced confounding, but the authors caution that it is not a general solution.
Jet model reports saturation-like scaling in heavy-ion jets
This preprint reports sizeable differences in illustrative RAA estimates between LHC-like and RHIC-like scenarios.
New construction disproves a broad bound in group theory
A preprint constructs countable torsion-free groups that violate the bound, while leaving the finitely generated case unresolved.
QCD framework links jet energy loss to color decoherence
Preprint: An arXiv study models how vacuum radiation and a quark–gluon plasma can be treated in one evolving picture.
Privacy noise helps protect data but can block recovery, study finds
Preprint: A mathematical analysis identifies a lower-noise regime for recovering a bounded hidden distribution and a square-root-scale regime where uniform consistency is impossible.
Sentence-level context tops a multimodal AI benchmark
An arXiv preprint reports its highest tested scores when visual elements are paired with reference-sentence context and document-level global context.
Strong Guide Fields Point to a Different Route for Particle Acceleration
Preprint simulations link approximately log-normal particle energies to charge-starved current sheets in relativistic plasma turbulence.
New Geometry Test Identifies When Tropical Multidegrees Are Positive
Preprint: A conditional projection test identifies positive multidegrees, while augmented Bergman fans also have Lorentzian volume polynomials.
Nonnegative Concave Functions Preserve Lee-Type Schatten Constants
This preprint reports an exact transfer for finite complex matrix families across 1 ≤ p ≤ ∞; the linear constants remain unknown specifically for 1 < p < 2.
New Muon variant uses fewer tokens in Transformer tests
An arXiv Preprint reports lower final validation loss and 13.3%–24.0% token-efficiency gains for SAMuon over tuned Muon in Transformer tests.
Human readers and language models track language trouble differently
Preprint: Conceptual disruptions showed a sharper, shorter pattern, while referential disruptions were more distributed across later reading and model prediction.
One Mathematical Framework Brings Several Plasticity Rules Together
Preprint: A single scalar functional supplies stresses, internal-variable forces, resistance and plastic-flow conditions in a theoretical framework.
AI model posts lower 50-second errors in airport forecasts
A version 2 arXiv preprint reports lower 50-second displacement errors for safety-critical and randomly selected agents on the Amelia-10 benchmark.
Preprint maps how genome length relates to species splitting
A mathematical model predicts distinct critical-length regimes, while finite simulations support the trends without testing them in organisms.