The information bottleneck (IB) trades compression of an input $X$ against prediction of a target $Y$ through a multiplier $β=1/λ$, yet each operating point is usually served by training, or fully con...
The paper investigates the limits of predicting how many tokens a large language model will generate using only prefill-phase hidden states. It shows that current benchmarks mix predictor quality with task difficulty, leading to unstable accuracy when models, datasets, or sampling settings change. By modeling decoding as a stopped stochastic process, the authors derive bounds and an affine variance law that explain the observed normality of per‑prompt decode lengths and confirm that system error dominates model error.
arXiv:2606. 16923v1 Announce Type: new Abstract: Simulation-based inference (SBI) of latent parameters is often hindered by simulator misspecification, the mismatch between simulated and real-world observations caused by inherent modeling simplifications.
The paper introduces a two-way guidance framework that jointly learns summary statistics and posterior distributions for Bayesian simulation-based inference. By imposing a structural restriction on the learned summary, the method ensures that the posterior remains sufficient, while the posterior objective simultaneously steers the summary toward sufficiency. The authors provide theoretical guarantees on posterior estimation error and the deviation of the learned summary from sufficiency, outlining conditions on simulation budgets needed to recover increasingly concentrated posteriors as observation size grows.
arXiv:2604. 16689v2 Announce Type: replace Abstract: Masking-based post-hoc explanation methods, such as KernelSHAP and LIME, estimate local feature importance by querying a black-box model under randomized perturbations.
arXiv:2610.03480v1 Announce Type: new Abstract: Post-training a large language model (LLM) often requires exploring trade-offs between multiple rewards, but retraining for each trade-off is expensive...
arXiv:2606. 30372v1 Announce Type: new Abstract: Quantitative research across the social and behavioral sciences depends on human subject experiments that are expensive, slow, and subject to sampling bias.
arXiv:2606. 10706v1 Announce Type: cross Abstract: Resource constraints increasingly determine what can be trained, fine-tuned, and deployed in large language models (LLMs), yet efficiency is often studied through isolated techniques rather than as an interacting system of limits.
arXiv:2505. 23869v4 Announce Type: replace-cross Abstract: A proposition that connects randomness and compression is put forward via Gibbs entropy over set of measurement vectors associated with a compression process.
arXiv:2610.09886v1 Announce Type: cross Abstract: Bayesian neural networks need not be fully stochastic to be universal conditional density approximators, but it remains open which parameters should...
arXiv:2606. 15871v1 Announce Type: cross Abstract: Bayesian inference for inverse problems is run to evaluate integrals -- posterior expectations, tail probabilities, and risks -- across a stream of observations.
arXiv:2610. 29425v1 Announce Type: new Abstract: Understanding generalization of randomized learning algorithms remains a central challenge.