Daily report — 2026-10-06

Spike-and-slab regression is a standard Bayesian formulation of variable selection: it returns a posterior distribution over which candidate effects are active rather than a single selected subset, so that every candidate effect carries an inclusion probability. Its cost grows exponentially with the number of candidate effects, so the posterior can be enumerated exactly only when the number of pre

Claimed: AUC 1.0

Evidence: - pretraining ("pretrained"); + cheap/closed-form ("closed form"); + toy/synthetic experiments ("synthetic"); + small/synthetic data ("synthetic"); + concrete method ("closed form"); + code: https://github.com/schiekiera/LinearPFN; + keywords: monte carlo

heuristic

We present the Prior-Fitted Language Model (PFLM), a 300M-parameter byte-level transformer pretrained only on samples from a synthetic non-linguistic prior. Given a prefix of real text, it learns to predict the language in context with frozen weights, having never seen a word of any real language.

Claimed: none extracted

Evidence: - pretraining ("pretrained"); + toy/synthetic experiments ("synthetic"); + small/synthetic data ("synthetic"); + concrete method ("update rule"); + formal statement ("convergence"); + claims simplicity ("simple"); + code: https://github.com/cbl/prior-fitted-language-model; + 2 HF upvotes

heuristic

Monitoring concept drift from an adaptive classifier's error stream creates an operational conflict with the model's own update loop. When internal adaptation outpaces evidence accumulation, accuracy recovers before cumulative detectors (CUSUM, Page-Hinkley) can reach threshold.

Claimed: none extracted

Evidence: + toy/synthetic experiments ("synthetic"); + small/synthetic data ("synthetic"); + public benchmarks ("benchmarks"); + formal statement ("bound"); + code: https://github.com/TheBlindSpotParadox/TheBlindSpotParadox-Experiments

heuristic

Tabular foundation models perform in-context learning (ICL) by conditioning predictions on labeled training examples provided as context. Unlike traditional models that separate training from inference, these models must process all training examples in every forward pass, making each prediction expensive.

Claimed: none extracted

Evidence: - LLM-based ("foundation model"); + toy/synthetic experiments ("synthetic"); + CPU-friendly ("lightweight"); + small/synthetic data ("synthetic"); + public benchmarks ("benchmark"); + formal statement ("bound"); + code: https://github.com/yoel-zeldes/tabalign

heuristic

We introduce Heuristic-Guided Fourier Fine-Tuning (HeuFouFT), a task-guided framework for selecting trainable frequency coordinates in Fourier fine-tuning. Existing uniform and Gaussian band-pass schemes allocate a limited spectral budget through fixed, task-agnostic rules.

Claimed: none extracted

Evidence: + toy/synthetic experiments ("Simulated"); + CPU-friendly ("lightweight"); + small/synthetic data ("Simulated"); + formal statement ("convergence"); - system/framework (many moving parts) ("framework"); no public code found; + keywords: optimizer, sampling

heuristic

Inference-time steering offers cheap, fine-grained control over a language model's outputs by estimating a concept's representation in activation space and shifting activations towards it. Existing methods build these representations from activation averages over contrastive datasets.

Claimed: none extracted

Evidence: + concrete method ("closed-form"); + formal statement ("theorem"); + claims simplicity ("simple"); + code: https://github.com/JDIXON-2/Concept_Activation_Transport; + keywords: ranking

heuristic

A language model can give a correct answer more probability than any single incorrect answer and still usually sample an incorrect one, because the incorrect answers together hold more probability. The power distribution raises each complete answer's probability to a power above one and renormalizes, shifting probability toward answers the model finds most likely (sharpening).

Claimed: accuracy 23.0; accuracy 5.3

Evidence: + formal statement ("Theorem"); + code: https://github.com/ArminAzizi98/OPPD; + keywords: sampling, monte carlo

heuristic

In decentralized consumer-to-consumer (C2C) marketplaces, people list goods, negotiate with strangers, and rate one another, so trust rests on reputation. Large language model (LLM) agents now act for users, raising risks to their money, privacy, and reputation.

Claimed: none extracted

Evidence: - LLM-based ("LLM"); + toy/synthetic experiments ("simulated"); + small/synthetic data ("simulated"); + public benchmarks ("benchmark"); + code: https://github.com/ziyan-wang98/BazaarBench

heuristic

We construct computable error estimators, which double as loss functions for neural networks, for a class of parametric nonlinear partial differential equations with a monotonicity property, and prove that they are globally reliable and efficient. The value of such a loss function is bounded above and below by the squared error in the natural trial norm, for every trial function, not merely for th

Claimed: none extracted

Evidence: + cheap/closed-form ("closed form"); + concrete method ("closed form"); + formal statement ("bound"); no public code found; + keywords: variational

heuristic

ufakzeka-karar is an open Turkish decision model with 182,494,466 parameters. Given a Turkish text and questions of a fixed answer type (a choice, a level on an ordered scale, or yes or no), it returns a temperature-scaled probability for every option and an expected error that serves as a "not sure" signal, without generating text and in one CPU forward pass for up to ten options.

Claimed: none extracted

Evidence: + CPU-friendly ("CPU"); + public benchmarks ("benchmark"); + code: https://github.com/ufakai/ufakzeka-karar

heuristic