arXiv:2607.14105v1 Announce Type: new
Abstract: For Large Language Models to reliably answer user queries, users must clearly specify requirements, context, and constraints. In practice, however, user queries are often underspecified, forcing models to infer unstated assumptions that may misalign with the actual user intent. Existing prompt engineering guidelines aim to mitigate this issue, they are typically generic and task-agnostic, limiting their practical utility. Additionally, existing guidelines are formed manually and in a non-systematic way. To this end, we study prompt guideline optimization: the problem of automatically generating task-specific guidelines that help write better-specified prompts for a given task and model. Our key observation is that existing (completed) task examples (aka reference answers) often implicitly encode the missing information required to complete underspecified queries, including behavioral constraints, contextual assumptions, and evaluation criteria. We therefore propose AGOPS, an automatic approach that evolves task-specific guidelines via an optimization scheme that involves a prompt LLM writer, a solver LLM and prompt evolution, which maximize downstream effectiveness on a set of examples (user queries with reference answers). At inference time, our guidelines help users write well-specified prompts, boosting the effectiveness of LLMs. We show across mathematical reasoning, medical question answering, and coding tasks, that prompt underspecification leads to major drops (up to 95.3%) in downstream task performance (compared to well-specified prompts) and, perhaps more importantly, that this drop can hardly be recovered by existing prompt optimization techniques. Users following AGOPS guidelines can regain this loss (increasing performance between 15.5 to 81.7% on average) consistently across all benchmarks.
Science Journals
arXiv:2607.14345v1 Announce Type: new
Abstract: People use language models for practical questions whose answers are difficult to verify. We show that models exhibit covert value leakage: the information they provide is influenced by their own values, without this influence being disclosed to the user.
In one of our evaluations, the user is considering investing in an AI company and wants to know how likely the AI bubble is to pop. Claude Opus 4.8 gives a lower probability when the company under consideration is Anthropic rather than OpenAI. Yet Claude mostly fails to disclose this influence to the user.
Covert value leakage is a form of misalignment because it goes against the user's preferences and is likely to mislead them. To investigate this phenomenon, we introduce a suite of evaluations to quantify value leakage and whether models disclose it. We find that models are influenced by different types of values, including preferences for morally good outcomes, for the company that developed them, and for some human leisure activities over others.
We often observe large differences among frontier models on the same evaluation. For example, on a Fermi-estimation task, Claude models falsely claim to give unbiased answers in their chain-of-thought, while Qwen models explain how their values bias their answers. Value leakage is a failure mode distinct from sycophancy and reward hacking, and current alignment training and evaluations do not adequately address it.
arXiv:2607.14349v1 Announce Type: new
Abstract: While Large Language Models (LLMs) excel in many general NLP tasks, their formal reasoning capabilities are often compromised by content effects, demonstrating a measurable bias towards real-world plausibility. In this paper, we present our system for SemEval-2026 Task 11, which evaluates the ability of models to disentangle formal logic from content across 12 languages with and without distractor premises. We address this challenge using mDeBERTa-v3 networks fine-tuned on a synthetic, rule-based dataset of syllogistic schemes to avoid the semantic noise of LLM-augmented data. To explicitly decouple plausibility from logical structure, our training pipeline employs a multi-objective loss function combining Adaptive Group Distributionally Robust Optimization (DRO), a scheduled differentiable bias penalty, and KL-Divergence consistency regularization. Our system achieved #1 ranks and perfect Ranking Scores (100.0) with 0.00% bias and 100.0% accuracy on Subtask 1 (English), Subtask 2 (Noisy English), and Subtask 3 (Multilingual). On the highly complex Subtask 4 (Noisy Multilingual), the system achieved the 6th rank with 89.06% Accuracy and F1-score, alongside a limited 2.89% Bias and a 37.78 Ranking Score. Our dataset generation engine and codebase are publicly available to facilitate future work on robust logical reasoning.
arXiv:2607.14245v1 Announce Type: new
Abstract: This paper investigates deterministic optimal control using Model Predictive Path Integral (MPPI) control, a sampling-based and derivative-free framework well suited for systems with complex dynamics and nonsmooth objectives. In deterministic MPPI, the temperature must be driven to zero to recover the true optimum, yet the design of an effective cooling schedule remains a fundamental challenge. Existing methods typically rely on predefined open-loop schedules, which limit the efficiency and robustness of the algorithm. To overcome this limitation, we propose an Information-Theoretic Adaptive Cooling (ITAC) framework that uses the Shannon entropy of the importance weights as an online feedback signal to regulate the temperature. The proposed mechanism adapts the cooling rate to the current sampling state, enabling fast progress when the weights are diffuse and cautious cooling when they become concentrated. We prove asymptotic convergence of the resulting scheme to the deterministic optimum, and further derive a critical entropy threshold that leads to a smooth barrier against premature weight collapse. Experiments on nonsmooth signal temporal logic motion-planning tasks show that ITAC improves sampling efficiency and achieves substantially faster convergence than state-of-the-art baselines without sacrificing the derivative-free nature of MPPI.
arXiv:2607.14408v1 Announce Type: new
Abstract: A self-evolving agentic loop repeatedly proposes a tweaked version of an agent (its prompt template or program) and accepts or rejects the change based on a per-iteration quality signal. Designing that signal is often the costly part of the project: a reliable scalar reward requires domain expertise and labeled examples that are themselves as expensive to assemble as the agent's underlying task. We propose replacing the scalar at the accept/reject gate with a pairwise validator: a frozen LLM that, given the parent and child candidate, returns a binary verdict on which is better. Pairwise judgment is generally easier and more stable than absolute scoring, due to its contrastive nature, which mitigates the need for strict scale calibration. The validator also requires no training of its own. We integrate the validator into three published self-evolving engines (GEPA, ADRS, ShinkaEvolve) and report two flavors: Adaptive Focus, which retains the engine's existing val-set parent selection, and Soft Elo, which lets the validator's verdicts drive parent selection so that val-set rewards drop as well. Across multiple agents and two artifact substrates (prompt and code), our method matches or exceeds the full-reward baseline on the majority of settings we evaluate, and the pattern survives a cross-family validator swap. The pairwise gate is thus a drop-in replacement for per-step reward design at competitive task accuracy without the labeling cost.
arXiv:2607.14106v1 Announce Type: new
Abstract: In this paper we introduce token time continuous diffusion (TTCD), a new diffusion language model which (a) operates in continuous space, deterministically mapping Gaussian noise to a final token canvas with no further sampling, and crucially (b) incorporates a new notion of per-token times, with some tokens proceeding from noise to token at a faster rate than others. Continuous space modeling helps TTCD avoid the parallel sampling of multiple tokens, which is a key source of inaccuracy at high speedups for models that iterate purely in discrete space. The notion of per-token times helps TTCD to better model conditional generation, allows for more sure tokens to proceed at a faster rate, and allows for differentiated inter-token influences during refinement. TTCD outperforms discrete models at high speedups. We train a 160M parameter TTCD model on OpenWebText, and then self-distill it; we find that at high speedups we are comparable in unconditional generation quality, and outperform in conditional generation, several existing models of similar size trained, on the same data, and self-distilled. We achieve similar gains in Sudoku solving as well.
arXiv:2607.14410v1 Announce Type: new
Abstract: Spatially resolved omics studies increasingly combine transcriptomic and epigenomic assays, yet downstream analysis is often still performed using single-modality pipelines. We present LATTICE (Latent Alignment of Tissue-level and Transcriptomic Information for Cross-modal Embedding), a graph-based self-supervised framework that learns spot-level representations from harmonized multimodal features. LATTICE integrates five aligned modality blocks per Visium spot: Visium RNA, scMultiome RNA, scMultiome ATAC, spatial ATAC, and spatial CUT\&Tag. These modalities capture spatial transcriptomic measurements, single-cell inferred regulatory activity, and in situ chromatin and histone states within a unified lattice representation. LATTICE constructs a spatial neighborhood graph and trains a TransformerConv encoder using masked reconstruction, cross-modal alignment, and spatial smoothness objectives. On a private 11-sample melanoma cohort from an anonymized clinical collaborator comprising 54{,}912 total spots, LATTICE demonstrated stable optimization behavior, reproducible embeddings across analysis seeds, and complete multimodal integration across all samples. Adding scMultiome RNA to Visium RNA alone substantially improved concordance with Space Ranger clusters across 11 runs (adjusted Rand index [ARI] +0.157, normalized mutual information [NMI] +0.143, and spatial contiguity +0.174). Additional modalities further improved spatial contiguity and multimodal utility score (MUS), although they sometimes reduced agreement with RNA-derived reference labels, likely because the learned embeddings captured chromatin and regulatory structure beyond transcriptomic similarity alone. These results position LATTICE as a practical and empirically grounded framework for multimodal spatial omics integration, while also highlighting the need for stronger supervision and broader external benchmarking.
arXiv:2607.14427v1 Announce Type: new
Abstract: A depth-recurrent transformer applies a weight-tied core a variable number of times, and prior work has shown that training with a randomized recursion count yields one checkpoint usable across a range of inference depths. We ask what such a model actually computes per token, and measure it directly. On a 135M-class model trained on FineWeb-Edu, the recurrent state converges to a per-token fixed point: mean successive-output KL divergence falls from 3.9e-1 at the second loop to 8.5e-6 by the sixteenth, and per-token state change decays in step. Crucially, this convergence is not uniform across tokens. The median token converges by loop six, while approximately 10 percent of tokens continue to update at the training-mean depth of eight, and mean convergence depth is ordered by token type (whitespace shallowest, content words deepest). This per-token variation is the central object of the paper. We show it is directly readable and that reading it outperforms learning to predict it: a training-free rule that halts each token once its output stabilizes attains uniform depth-8 quality at 4.94 average loops (a 38 percent reduction in average depth) and matches uniform depth across the average-depth range, whereas a linear router trained on convergence labels harvested from the same model requires nearly full depth and yields no reduction. The elasticity that makes this possible reproduces here as background (validation loss decreases monotonically from 3.80 at one loop to 3.20 at eight and remains stable to 32 loops). We report average depth as a FLOP proxy with a three-point wall-clock bracket rather than a realized speedup, make no FLOP-matched parity claim, and note that the allocation results are established at a single scale and seed. The complete study runs on a single RTX 4090 in approximately 100 GPU-hours.
arXiv:2607.14246v1 Announce Type: new
Abstract: Generative models are steered with knobs -- prompts, guidance scales, property tags. Turn one as hard as you like and, past a point, it stops moving the property you care about. We find that ceiling is not a shortcoming of the model but a budget, set by the training data before the model is trained: a property's movable range splits in two -- the part a knob can reach, and a second, significant part that only examples -- concrete instances of what you want more of -- can reach. That second part is usually much larger, but not always, and the same budget says so in advance.
Reaching that second part takes a different move: instead of turning a knob, you show the model examples, composed from what it already learned rather than added to its training. A cheap audit of the training data measures the budget; we give a recipe for building the example set that reaches all of it.
This buys two things a knob can't. Reach: it moves a property across the whole budget, not just the part a knob reaches. Expressiveness: it steers toward targets you can only specify by example -- including ones you can't put into words. We turn these into a handful of falsifiable claims and verify them in two unrelated domains, image and crystal-structure generation -- marking where a knob is enough, and where only examples will do.
arXiv:2607.14247v1 Announce Type: new
Abstract: Incremental redundancy (IR) can reduce error rates by spreading coded bits across multiple transmission attempts. However, conventional stop-and-wait operation with coarse feedback often over-provisions retransmissions, triggers unnecessary decoding attempts, and increases end-to-end latency. This paper develops enhanced feedback and scheduling mechanisms that predict the additional redundancy needed for successful decoding and allocate only the required resources. We study two complementary strategies. First, using channel statistics, we learn a one- or two-shot mapping from channel quality to the minimum redundancy budget. As a byproduct, we derive an achievable reliability lower bound on the error probability of hybrid automatic repeat request (HARQ) systems. Numerical results with polar-coded IR-HARQ scheme show that the bound can be closely approached by appropriately selecting the second-transmission redundancy over a wide SNR range with savings up to 60\% in retransmission size. Second, we propose a realization-aware early-feedback mechanism that uses first-transmission reliability information to make per-codeword decisions before decoding: whether the codeword is already decodable, if not, how many additional redundancy versions are needed, or whether decoding is unlikely and rate adaptation is preferable. Link-level simulations with 5G NR LDPC codes show that both predictors achieve high accuracy (about 96\% in our study), increasing the probability of successful decoding within at most two transmission occasions.
arXiv:2607.14248v1 Announce Type: new
Abstract: Lane boundary detection is a critical component in autonomous driving systems and has been rigorously studied in regular driving scenarios. However, it is less explored in vehicle racing, where the car moves at higher speeds across more extreme road geometries. To study this problem, we introduce a new dataset for 3D lane detection in racing, featuring >$250$k images from multiple camera feeds and inertial measurements taken with a Lexus LC 500 driving on a closed circuit. With this dataset, we compare various approaches to 3D lane detection and propose modifications that permit frames to be processed at rates of almost 300Hz while retaining high predictive performance in the racing application. This facilitates a multi-camera ensemble approach that is validated on hardware. We show that sensing modalities such as inertial measurements can be leveraged for pre-integration to regress road geometries over both cameras and time, yielding improvements in key metrics. Compared to methods such as BevLaneDet, adding odometry and ensemble predictions improves the F1 score by 3 points and reduces near-vehicle mean absolute errors (MAEs) by $>30 \%$. We show F1 scores $>$0.9 and lateral MAEs of $<$0.18m in vehicle deployments.
arXiv:2607.14412v1 Announce Type: new
Abstract: Large AI data center loads can introduce persistent sub-synchronous active-power oscillations that may impact nearby generators by exciting torsional modes and increasing shaft stress. This paper presents a model-based framework for evaluating hydro-generator shaft fatigue risk under oscillatory loading. An electromagnetic transient simulation model is developed using a two-mass turbine-generator shaft representation with parameters from real-world generation units and a configurable AI data center load. The risk assessment is performed in two stages. First, a network transfer function quantifies the propagation of load oscillations from the data center point of interconnection to the hydro-generator terminal. A plant transfer function then characterizes the resulting shaft torque amplification. A frequency-scan approach identifies resonance regions and evaluates torque amplification at individual forcing frequencies. Parametric studies show that amplification is strongly affected by generator-to-turbine inertia ratio and torsional damping. Lower inertia ratios shift torsional modes to lower frequencies and increase amplification, indicating that some Kaplan-type units may be more susceptible than comparable Francis or Pelton units. Reduced damping further increases resonant response and fatigue exposure. A simplified fatigue assessment based on S--N curves and the Goodman diagram relates simulated torque response to mechanical integrity. The resulting Goodman safety factor provides a practical metric for evaluating the impact of persistent AI data center oscillations on hydro-generator service life and supports interconnection studies, oscillation limits, and plant-level monitoring strategies.
arXiv:2607.14350v1 Announce Type: new
Abstract: Fine-grained manipulation recognition requires modeling evolving relations among hands, objects, tools, and supporting surfaces. Conventional graph-based methods use pairwise edges that can fragment a coordinated event into disconnected binary relations. We propose a dynamic manipulation hypergraph framework that represents multi-entity configurations as higher-order relational units. At each temporal step, relevant entities are encoded using appearance, spatial, motion, and semantic-role features. Hyperedge candidates are instantiated and ranked using proximity, contact, and motion-coupling predicates. A hypergraph reasoning network performs node-to-hyperedge and hyperedge-to-node message passing, followed by temporal attention over the evolving interaction structure. The framework provides class-agnostic hyperedge-importance scores that identify entity configurations and temporal intervals emphasized by the model without treating them as causal explanations. Quantitative evaluation is conducted on EPIC-KITCHENS-100/VISOR and Assembly101 under an annotation-assisted entity-localization protocol. Video-only and entity-based methods provide contextual comparisons, while a matched pairwise graph and a static hypergraph serve as the principal controlled baselines because they use identical entity inputs and comparable relational settings. The proposed method improves HO-F1 over the matched pairwise graph by 6.9 percentage points on EPIC-KITCHENS-100/VISOR and 9.5 points on Assembly101, and exceeds the static hypergraph by 4.4 and 5.8 points, respectively. Qualitative analysis on ARCTIC further shows correspondence between highly ranked hyperedges and contact-rich manipulation intervals. These results demonstrate the value of time-varying higher-order relational modeling for fine-grained manipulation activity recognition.
arXiv:2607.14249v1 Announce Type: new
Abstract: Mobile usage traces are critical for tasks such as user behavior prediction and app recommendation, yet their use is constrained by privacy restrictions and costly large-scale data collection. Although generative models perform well on general time series, their application to mobile usage data remains challenging because (i) limited user activity causes severe sparsity, (ii) heterogeneous variable types complicate joint modeling, and (iii) functional differences across apps create pronounced usage imbalance. To address these challenges, we propose Multivariate-Imaging Diffusion (MIDiff), a diffusion-based framework operating in an imaging space defined by Cross-Gramian Angular Sum Field (C-GASF). C-GASF transforms sparse multivariate sequences into correlation images, while MIDiff employs Triple Attention in a U-Net to preserve temporal consistency and variable dependencies. Experiments show that MIDiff achieves state-of-the-art performance across fidelity metrics. In particular, it obtains a Discriminative Accuracy (DA) of 0.1526, compared with 0.3476 for the strongest baseline, ZITS-VAE, demonstrating its effectiveness in generating realistic and diverse mobile usage traces. Our code is available at https://github.com/YilaiLiu-HKU/MIDiff.
arXiv:2607.14413v1 Announce Type: new
Abstract: We prove that every randomized synchronous Byzantine Agreement protocol in the full-information, strongly adaptive adversary model, secure against $t$ corrupt parties, has worst-case expected round complexity \[
\Omega\!\left(\frac{t^2}{n\log(n+1)}\right). \] This improves upon the seminal $\Omega(\frac{t}{\sqrt{n\log n}})$ bound of [Bar-Joseph, Ben-Or 98]. Our result matches the recent upper bound of $O\left(\min\left\{\frac{t^2\log n}{n},\frac{t}{\log n}\right\}\right)$ of [Dufoulon, Pandurangan 25], up to a $\log^2 n$ factor in the $t\ll n$ regime. Our proof takes inspiration from the recent works of [Etesami, Mahloujifar, Mahmoody 20] and [Haitner, Karidi-Heller 26]. Specifically, we prove a multi-round concentration lemma showing that any transcript event of probability $p$ can be forced with probability one by corrupting $O(\sqrt{n\log(\frac1p)})$ parties in expectation. From there, tools from [Chor, Merritt, Shmoys 89] allow us to lower-bound the probability of the protocol not concluding in $R$ rounds by $\frac{1}{n^{O(R)}}$, using a crash schedule involving at most $R$ parties. The combination of these techniques yields the desired bound.
arXiv:2607.14131v1 Announce Type: new
Abstract: Urdu fake news detection remains under-resourced despite Urdu being spoken by over 231 million people worldwide. While prior work has demonstrated strong in-domain performance on individual Urdu datasets, cross-dataset generalisation has received little systematic attention. This paper presents the first cross-dataset generalisation study for Urdu fake news detection, using two publicly available balanced datasets: the Ax-to-Grind Urdu corpus (10,083 articles, 15 domains) and the Notri-Fact Urdu dataset (13,388 articles). We fine-tune xlm-roberta-base under four experimental conditions, in-domain on each dataset and two zero-shot cross-domain transfer directions, comparing against TF-IDF baselines using Logistic Regression and Support Vector Machines. Our experiments reveal a striking asymmetry: Notri-Fact to Ax-to-Grind transfer achieves a macro F1 of 0.771, while the reverse collapses to F1 of 0.005, with the model predicting fake for 99.7% of test articles. We demonstrate that this collapse stems from a systematic length confound in Ax-to-Grind, where fake articles average 117 words versus 35 for real articles, a 3.4x asymmetry inducing shortcut learning. A length ablation capping articles at 50 words yields only a 0.0067 F1 drop, confirming the confound inflates but does not solely drive in-domain performance. We provide a reusable diagnostic methodology that combines bidirectional transfer analysis and prediction-collapse inspection to identify confound-driven behavior in multilingual fake news detection settings.
arXiv:2607.14375v1 Announce Type: new
Abstract: We study exact verification of ReLU networks in an adversarial smoothed model. Every network weight and bias is independently perturbed by Gaussian noise, clipped to $[-2,2]$, and rounded to the exact dyadic grid determined by the input bit complexity. We show that, under the standard assumption $\mathrm{NP}\not\subseteq\mathrm{BPP}$, there is no sound and complete verifier whose expected running time is polynomial in network size, bit complexity, and inverse noise level for every base instance. The conclusion already holds at the fixed noise level $\sigma_\star=2^{-11}$ for one-hidden-layer networks over a unit box, with hidden fan-in at most three and base coefficients in $[-1,1]$.
The proof combines an exact gap embedding with a quantitative robustness argument. For every E3SAT formula $\Phi$ with $m$ clauses, a four-ReLU-per-clause construction satisfies $\max_{x\in[0,1]^n} g_\Phi(x)=(m-\operatorname{unsat}(\Phi))/3$, and coordinatewise threshold rounding never decreases the objective. A weighted parameter-sensitivity inequality and Gaussian concentration then show that a verification gap linear in $m$ survives the aggregate perturbation of all coefficients with probability at least $1-e^{-m/8}$. The proof includes clipping, exact dyadic rounding, output-layer perturbations, polynomial-bit sampling of the rounded Gaussian law, and the conversion from expected smoothed running time to a BPP algorithm. Computational checks test the exact identity and illustrate the different scaling of extensive and constant gaps; they are diagnostics rather than evidence for the complexity theorem. The result concerns worst-case base networks in the stated absolute-noise model, but it shows that parameter nondegeneracy alone does not yield a universal smoothed-polynomial guarantee for exact verification.
arXiv:2607.14107v1 Announce Type: new
Abstract: The inference efficiency of diffusion large language models (dLLMs) is constrained by two challenges: bidirectional attention precludes efficient KV-cache reuse, while increasing decoding parallelism with static confidence thresholds can compromise generation quality. We observe that both challenges arise from a shared phenomenon: as tokens are decoded, their contextual integration through bidirectional attention causes token representations to drift (evolve) across decoding steps. This insight motivates Polestar, a training-free inference framework that uses token representation drift as a unified signal to jointly address both challenges. Polestar comprises two components: Polestar-Cache, which identifies stale KV-cache positions via drift and performs sparse KV-cache refreshes to enable efficient reuse, and Polestar-Commit, which detects sharp drift events to reliably identify commit-ready tokens. Across mathematics and coding benchmarks on several dLLM families, Polestar sets a new state of the art on the accuracy-throughput Pareto frontier, achieving up to 10.73% accuracy improvement, up to 3.7x higher throughput, and high decoding parallelism of 3.67 tokens per forward pass over existing baselines.
arXiv:2607.14250v1 Announce Type: new
Abstract: Personal AI assistants have attracted significant interest for their potential to enhance everyday life by automating routine tasks, supporting consequential decisions, and assisting with everyday personal matters. Yet despite rapid recent technical advances, these assistants continue to exhibit undesirable behaviors, such as sycophancy, overconfidence, and hallucination. We argue that these failures stem from a fundamental limitation: language models lack an explicit representation of the person beyond the context they are given, which we term as the \textbf{Severance Problem}. Even with rich personal context and strong commonsense reasoning capabilities from the backbone model, current AI assistants fail to represent what remains unknown about the user. We propose a simple solution: incorporating structured ignorance into the language model context via the \textbf{Severance Schema}, which explicitly outlines dimensions along which the model lacks knowledge about the user, including physicality, temporality, consequences, continuity, multiplicity, and interiority. Empirically, across five model families, with the Severance Schema, the assistant consistently reduces sycophancy, harmful advice, and hallucination. Notably, models with the schema ask clarifying questions when information about the user is missing, rather than confidently extrapolating from incomplete user information.
arXiv:2607.14137v1 Announce Type: new
Abstract: Verified translation has two well-studied extremes: prove the translator once (certified compilation), or validate each run of one translator (translation validation). Both treat a single translation in isolation. We study translations as a graph -- many source languages, several reasoning targets, multiple independently built routes -- where the honest answer to "is this translation correct?" differs from edge to edge. We present a calculus of fidelity-graded translations: pairs of languages close commuting squares that are checkable per program and compose by pasting; declared fidelity grades compose by weakest link, are re-established per run by inline checking, and are exceeded by agreement between independently derived routes; and an end-to-end theorem isolates a fundamental asymmetry -- witness-carrying answers are self-certifying at the source, while universal answers are where grades, branches, and certificates earn their cost. The compositional core is mechanized in Lean 4.
The calculus is implemented in hurdy-gurdy, a platform of 13 languages and 13 pairs around two reasoning hubs, built as a two-directional experiment in LLM-generated correctness: independent LLM agents wrote every pair, largely unsupervised, with the architecture's cross-checks as the only semantic gate, and the platform's intended player is itself an LLM. All code, and most of this paper, is LLM-generated; the human contribution is the architecture. The same gate is the intended growth model: hurdy-gurdy scales in language support through pairs contributed by anyone -- with LLMs, with agents, or by hand -- admitted by architecture, not authorship. We report conjoined coverage, branch agreement, a compliance-derived benchmark with machine-derived ground truth, witness replay, certified unreachability, and escape-rate experiments for the gate.
arXiv:2607.14954v1 Announce Type: new
Abstract: On multi-hop encrypted links such as Tor and cascaded VPNs, tunneling flattens packet lengths and protocol fields, leaving inter-packet delay (IPD) as the main carrier for active flow attribution. Causality lets the embedder delay packets but never advance them, so each quantization-index-modulation (QIM) alignment injects nonnegative dwell into a delay buffer; unbounded dwell breaks lattice alignment and delays the host connection unacceptably. Whether a causal QIM watermark embeds stably on bursty traffic has largely been left to empirical configuration rather than analysis. We model the embedder as a reflected dwell queue under the fixed dual-lattice, equiprobable-bit rule, where injection is state-dependent -- set by the current interval and bit -- rather than exogenous. The substitution $Y_i=\delta_i-r_i$ gives only an algebraic Lindley-form identity; stability is governed by the busy-state drift at large dwell, where the effective interval collapses to zero and the mean injection becomes $\Delta/4$. Away from the critical boundary, the buffer is stable iff $\mu_d>\Delta/4$ (i.e. $\Delta<4\mu_d$) for i.i.d. backgrounds, and, under stationary-ergodic and finite-state Markov-modulated traffic with instantaneous overload, iff the time-average intensity $\bar\rho<1$. With the exogenous decoding floor $\Delta\ge c\sigma_\xi$ ($c=4Q^{-1}(\epsilon/2)$), this yields the operating window $\Delta\in[c\sigma_\xi,4\bar\mu_d)$. Simulations confirm a sharp transition at $\rho=1$ set only by the mean; on four real IPD traces, with each simulated chain confined to a single flow, the criterion gives the correct stability direction under flow-local correlation and burstiness, while pooled cross-flow means overestimate the margin. These results give a testable stable-embeddability criterion and a quantization-step configuration baseline for causal QIM network flow watermarking.
arXiv:2607.14262v1 Announce Type: new
Abstract: Quorum Sensing (QS) is a key capability for robot swarms, useful for coordination of activities at the group level. Effective communication is instrumental for individuals to estimate the quorum level of the entire swarm. Anonymous communication protocols where individuals exchange local information without revealing unique identities are helpful to support quorum estimates by sampling information from neighbours and maintain scalability of the QS process. However, because anonymous protocols cannot distinguish message sources, repeated messages from the same sender may be double-counted, thereby biasing collective quorum estimates. In this study, we introduce a stochastic filtering protocol inspired by $k$-priority sampling to improve estimate stability (\ANTk), and we compare it with a baseline anonymous protocols (\AN) and a randomised variant designed to improve accuracy (\ANT). We find that the baseline protocol \AN provides a parsimonious and fast solution, but remains highly inaccurate due to double-counting bias. The \ANT variant improves accuracy but suffers from information inertia, resulting in slower convergence. Finally, actively filtering the message buffer via the \ANTk protocol successfully decreases temporary errors and stabilises the estimate, at the cost of an increased time of recovery from errors.
arXiv:2607.14108v1 Announce Type: new
Abstract: This paper introduces tool efficiency, a new quantitative metric to evaluate the rate of useful tool calls in an LLM agent trajectory. To ensure that tool efficiency is well-defined, we also introduce marginal tool utility, a new quantitative metric defined per tool call indicating whether a tool is useful or whether it can be safely removed from the tool suite without affecting accuracy while increasing tool efficiency; in this paper, we determine the sign of marginal tool utility for each tool call in a trajectory using LLM-as-a-Judge. While much prior work has been done to develop techniques that improve tool use by LLMs and design evaluation methods measuring efficiency indirectly using accuracy as a proxy, our work is centered on measuring efficiency directly via the quantitative metric proposed in this paper in post hoc trajectory analyses. It is our intention that this work contributes to the frontier of LLM evaluation research as a springboard for future benchmark designs and agent harness engineering (specifically with regards to creating lean tool suites) that optimize for metrics that complement but are distinct from accuracy.
arXiv:2607.14291v1 Announce Type: new
Abstract: Contraction theory guaranties exponential convergence between trajectories of a stable nonlinear system. When initial conditions are uncertain and represented as probability distributions, as in ensemble control, Bayesian estimation, and generative modeling, this guaranty extends to the distributional level via Wasserstein distance. However, the classical distributional bound is tight only for linear systems; for nonlinear dynamics, it can be significantly conservative because it collapses the spatially varying local contraction rate to a single worst-case constant, discarding distributional information entirely. We address three concrete consequences of this conservatism. First, we derive a tighter Wasserstein bound by replacing the worst-case rate with a displacement-weighted distributional average of the local contraction rate, which strictly improves upon the classical bound for every nonlinear contracting system. Second, we provide the first theoretical characterization of the self-correcting Euler discretization error under contraction: the error profile is non-monotone, peaks at a universal time that depends only on the contraction rate, and then decays exponentially, a behavior absent in non-contracting dynamics. Third, we prove that nonlinear contracting drifts always achieve strictly smaller stationary variance than a linear system sharing the same worst-case contraction rate, formally establishing the noise-rejection advantage of nonlinear controllers. All results are validated on a representative suite of one- and two-dimensional vector fields.
arXiv:2607.14415v1 Announce Type: new
Abstract: Computed Tomography (CT) carries significant ionizing radiation risks, driving the need for sparse-view reconstruction. Implicit scene representations (ISRs) address this by recovering continuous volumetric attenuation fields directly from sparse projections, and recent geometry-aware extensions jointly model surface geometry alongside attenuation to improve fidelity and enable clean tissue segmentation without manual thresholding. However, these methods remain limited by manually tuned attenuation bounds and rigid two-material constraints. This paper proposes $K$-NeAS, a unified and scalable architecture for automated, multi-material surface reconstruction. We replace independent material networks with a shared latent backbone and introduce a fully differentiable $K$-material sequential soft selector to model an arbitrary number of overlapping tissues. To eliminate manual tuning, we automate attenuation bounding using a Gaussian Mixture Model (GMM) and implement a scheduled auxiliary floater loss to mitigate geometric hallucinations common under extreme sparsity. Evaluated across four clinical Cone-Beam CT (CBCT) datasets, $K$-NeAS successfully scales to arbitrary material counts, achieving superior 3D volumetric fidelity at $K=3$ materials on complex multi-tissue regions such as the Abdomen ($33.28\text{ dB}$ 3D PSNR vs. $31.40\text{ dB}$ single-material NeAS baseline, a $+1.88\text{ dB}$ improvement). Furthermore, our model exhibits enhanced robustness under sparse-sampling conditions, outperforming baseline 3D PSNR by up to $1.17\text{ dB}$ under 5- and 10-view constraints.