Vision-language models (VLMs) are increasingly proposed as general-purpose tools for scientific data interpretation, yet their reliability on real astronomical observations across diverse modalities remains untested. We present AstroVLBench, a comprehensive benchmark comprising over 4,100 expert-verified instances across five tasks spanning optical imaging, radio interferometry, multi-wavelength photometry, time-domain light curves, and optical spectroscopy. Evaluating six frontier models, we find that performance is strongly modality-dependent: while one model (Gemini 3 Pro) emerges as the most consistently capable across tasks, task-specific strengths vary, and all models substantially underperform domain-specialized methods. Mechanistic ablations reveal that performance depends not only on directing attention to salient visual features but also on grounding those features in physical knowledge. Phenomenological prompts describing what to look for improve accuracy by sharpening model focus, but physical prompts explaining why those features matter perform better overall and yield more balanced classifications with reduced class-specific bias. Consistent with this picture, presenting the underlying one-dimensional measurements directly as numerical tables instead of rendered plots yields up to 13 percentage points improvement. Reasoning quality analysis further demonstrates that, without explicit physical grounding, models may reach correct predictions from phenomenologically plausible cues while providing physically imprecise justifications, establishing that accuracy alone is insufficient for trustworthy scientific deployment. These findings provide the first systematic, multi-modal baselines for VLMs in observational astronomy and identify the specific representation, grounding, and reasoning bottlenecks where current models fail.
Scattered light is one of the most common sources of non-stationary noise at low frequencies in Advanced LIGO detectors. It appears as arch-like features in time-frequency spectrograms, produced when stray light reflects from moving surfaces and recombines with the main interferometer beam. In this study, we present ArchGEM, an automated framework for identifying and characterizing these arches and recovering the physical properties of the scattering surfaces. ArchGEM combines a prominence-based peak-finding method with a Gaussian Mixture Model clustering approach to capture a range of scattered-light morphologies across different detector conditions. We apply ArchGEM to scattered light glitches across Advanced LIGO observing runs O3 (2019--2020) and O4 (2023--2024). We find that the average frequency distributions of this noise span 15--25 Hz in O3a and O4, but increase to 20--40 Hz during O3b. Typical inferred surface velocities are 0.2--0.5 $μ$m/s, and inferred surface displacements are 0.1--0.3 $μ$m. The Gaussian Mixture Model performs most consistently for complex or overlapping features, with mean frequency offsets within 5 Hz of the Gravity Spy baseline. Our results show that ArchGEM provides a practical tool for detector characterization by linking observed spectrogram features to the motion of scattering surfaces and helping guide future mitigation of scattered light noise in current and next-generation interferometers. By quantifying the temporal and spectral behavior of scattered light, ArchGEM provides a robust framework for diagnosing noise sources and guiding targeted mitigation strategies in future detector upgrades.
The quiet-Sun coronal electron-temperature ratio $R \equiv T_\mathrm{EUV}/T_B \approx 2.4$, stable across an eight-year solar cycle, is read here as a measurement of relative entropy between two diagnostic projections of the coronal electron distribution onto the one-parameter Maxwellian family. The EUV ionization temperature is a moment-matching projection against a Bethe-type ionization kernel; the radio brightness temperature is the Rayleigh-Jeans source function of thermal bremsstrahlung. For a kappa distribution in the mean-energy convention, Fleishman & Kuznetsov (2014) give the radio-side projection in closed form as $T_B = T_\mathrm{core}$; the EUV side returns $T_\mathrm{eff}$ up to a shape-dependent correction within the Dudík et al. (2014) intensity-ratio envelope. At $κ= 2.5$ the Kullback-Leibler divergences between the true distribution and its two Maxwellian projections evaluate to $0.32$ and $1.20$ nats, and their difference satisfies $ΔD_\mathrm{KL} = (3/2)[R_0 - \ln R_0 - 1] = (3/2) d_\mathrm{IS}(T_\mathrm{eff}, T_\mathrm{core})$, where $R_0 \equiv κ/(κ- 3/2)$ is the ideal closed-form ratio and $d_\mathrm{IS}$ is the Itakura-Saito distance. The identity is offered as an analytical reference for observational systems in which two diagnostics project different moments of a common non-equilibrium distribution; the eight-year stability of $R$ expresses a stability of that projection structure.