arXiv subjects

Paras Balani

Publications and source records attributed to Paras Balani.

3 recordsLinked to original sources

arXiv2026

LLMs Don't Pay for the Jump

Zahavy [2026] argues that Large Language Models, despite their capabilities in induction and deduction, cannot perform the abductive "Jump" that produced Einstein's equivalence principle, and attributes this limitation to the absence of embodied simulation. Zheng-Xin [2026] and Farmer [2026] question whether embodiment is necessary for abduction, pointing to alternative routes to General Relativity and forms of abduction that require no sensorimotor grounding. Max Planck resolved the blackbody radiation problem in 1900. Planck's move to E = h{\nu} required no embodied simulation. It was motivated by a mathematical consequence of classical theory, an infinite predicted energy for a finite measured quantity, that could not be physically accepted. We show that neither induction nor deduction could have produced the postulate and argue that its adoption required a coupling between epistemic error and physical cost. We formalize this distinction through thermodynamic coupling and show that fixed-weight transformer inference lacks such coupling, regardless of model scale. This is consistent with empirical results showing that output entropy remains nearly unchanged across tasks with sharply increasing causal difficulty, even as accuracy falls from 100% to 17%. We therefore argue that the missing ingredient in machine abduction may lie deeper than embodiment: a system must have some physical mechanism through which epistemic error becomes costly enough to force revision.

cs.AI↗

arXiv2026

Self-Referential Induction Increases Response Instability Relative to Unresolvable and Verifiable Questions in Large Language Models

Self-referential prompting has been shown to reliably induce large language models to produce first-person reports resembling subjective experience, but no prior work measures how consistent these reports are across repeated, independent trials, or how that consistency compares to the model's behavior on other kinds of open-ended questions. We measure response instability, defined as one minus the mean pairwise cosine similarity of sentence embeddings computed over a compressed core claim extracted from each response, for three groups of questions: self-referential prompts eliciting a subjective-experience report, unresolvable philosophical questions unrelated to self-reference, and questions with a verifiable correct answer. Using 30 independent responses per question (360 responses total, Gemini API, temperature 0.7) across four questions per group, we find that self-referential questions show the highest instability (0.343 +/- 0.047), unresolvable philosophy questions show intermediate and tightly clustered instability (0.192 +/- 0.008), and verifiable questions show the lowest instability (0.105 +/- 0.058). This provides a quantitative baseline for the induced subjective-experience report, showing that it occupies a distinct, less stable position in the model's output distribution than ordinary open-ended philosophical uncertainty.

cs.CL↗

arXiv2026

Observational Signatures of Static and Rotating Wormholes Embedded in Dark Matter

We study static and slowly rotating traversable wormholes embedded in two contrasting dark matter environments, the cuspy Navarro-Frenk-White (NFW) halo and the cored solitonic wave dark matter ($\psi$DM) profile, using observational parameters set by the rotation curve of the dwarf galaxy NGC\,2366. For each profile, we solve the Einstein field equations in the Morris-Thorne framework to obtain the shape and redshift functions, then extend to slow rotation via the Teo metric with a Lense--Thirring frame-dragging term. Across the four resulting spacetimes and six throat radii, we trace null geodesics, compute specific intensity profiles and polar shadow maps, and construct accretion disk images including the full relativistic Doppler effect. We noticed these two profiles differ sharply in photon dynamics. The NFW potential is too shallow to support a detached photon sphere, so its critical impact parameter stays close to the throat radius. The soliton core, by contrast, is focused enough to host a genuine unstable photon orbit that raises the critical impact parameter to about $1.18$--$1.26$ times the throat radius. This difference in photon-sphere structure propagates through nearly every observable, including photon-ring sharpness, shadow size, accretion-disk appearance, and, in the rotating case, the degree of shadow asymmetry from frame dragging. The soliton shadow size also tracks the mass of the underlying ultralight boson through $\rho_c\propto m_b^{-2}$, with a transition from total capture to mainly deflecting lensing near $m_b\sim10^{-18}$\,eV, a feature absent for NFW that offers a direct probe of dark matter microphysics. These results show that wormhole photon dynamics are shaped by the dark matter profile, pointing to a possible observational route to probe the nature of dark matter.

gr-qc↗