Records published in October 2026
37 records, all free to read and download.
All of 2026-
Journal article
A Rewrite That Keeps the Label Has Already Read It: Why an LLM Paraphrase Can Raise a Sentiment Classifier's Accuracy Without the Classifier Getting Better, Four Accounts of the Gain, and the Re-Annotation and Neutral-Rate Controls That Would Separate Them
A 2026 study reports that sentiment classifiers score higher on reviews rewritten by small language models than on the human originals, by up to 5.8 accuracy points, and reads the gain as stylistic alignment: the rewrite strips idiosyncrasy and hands the classifier text closer to what it…
-
Journal article
Who Spoke Next Is Not Who Should Have: What Zero-Shot Failure on Multi-Party Turn-Taking Benchmarks Measures, Why the Human Reference Sits Beside the Models, and What Would Show That Knowing When to Speak Must Be Trained
An assistant that joins a group conversation has to decide, at every pause, whether to take the floor. Two 2026 benchmarks report that language models prompted zero-shot make this decision badly and that fine-tuning repairs it, and both conclude that knowing when to speak is learned…
-
Journal article
A Blind Score Is Not a Reliance Measure: What Answering Without the Audio Shows About Audio-Language Benchmarks, Why the Partial-Input Debate Already Settled Half of It, and the Counterfactual-Audio Test the Other Half Needs
Audio-language models are now routinely tested with the audio taken away, and the results are read in opposite directions. One 2026 audit reports that eight models keep 60 to 72 percent of their accuracy on three benchmarks when the audio is omitted, and concludes that the scores largely…
-
Journal article
Which Signal Did the Synthesis Import? Information, Re-presentation and Selection in Synthetic Pretraining, Why 'the Verifier Is Doing the Work' Holds Only Where Correctness Is the Target, and the Controls That Would Assign the Credit
The expectation that language models will keep improving after the supply of human text runs out rests on a claim about synthetic data: that a model, or a family of models, can produce training text worth more than the text it was trained on. A competing reading holds that no…
-
Journal article
Robust Because Coarse, or Robust Because Planned? A Research Agenda for Training-Data Provenance After the Aggregate-Disclosure Turn, in Which Granularity Buys Statistical Power, Prior Randomisation Buys Soundness, and Per-Example Value Has Neither
Regulators have chosen coarse disclosure of training data: the template for the EU AI Act's public summary of training content asks for lists of large datasets, a summary of the most relevant crawled domains and a description of how opt-outs were honoured, and does not require a list of…
-
Journal article
Lossless at What Length? Near-Lossless KV Cache Compression Is a Claim Indexed by Decode Exposure, Scoring Statistic and Budget Accounting, the Decode-Aware Line Has Answered the First Index Mostly on Final-Answer Accuracy, and Divergence From the Full Cache Is Not by Itself the Defect
Key-value cache compression is routinely reported as near-lossless at aggressive budgets: eviction, quantisation and merging methods keep a fifth or a tenth of the cache and match the full-cache score. A 2025-2026 line of work argues that the result was manufactured by evaluation…
-
Journal article
INT4 Is Not One Treatment: Why the Quantisation-Safety Literature Cannot Be Reconciled as Reported, How Six Design Axes Separate Its Studies Before Bit-Width Does, and Why Two 2026 Studies of the Same Model Names Reach Opposite Verdicts
Post-training quantisation to 4 bits is widely treated as nearly free, a judgement resting on perplexity and multiple-choice accuracy. Whether it is also free for safety is now contested: 2025-2026 studies report that 4-bit checkpoints lose refusal while their quality holds, that 3-bit…
-
Journal article
Kuwait Gratuity Calculator 2026: A Practical Employee Guide
Learn how to use a Kuwait Gratuity Calculator to estimate end-of-service indemnity. Understand Article 51, salary requirements, service years, resignation rules, and final settlement calculations.
-
Journal article
Where the Transcript Runs Out: When Model Internals Beat Behavioural Monitors, Why the Strongest Generalisation Evidence Comes From Targets the Text Already Shows, and the Comparisons No Study Has Yet Run
Linear probes and other readers of a language model's activations are now deployed as safety monitors, and two published literatures disagree about what they buy. One reports probes that match prompted and fine-tuned language-model monitors on real-world data after training only on…
-
Journal article
What Counts as a Control? A Scaffold Effect Is a Contrast Conditional on What Was Held Fixed, the 2026 Controlled Studies Each Held Something Different, and Only a Declared Spend Interval Makes a Harness Claim Refutable
A language-model agent is a model wrapped in a harness that builds its context, exposes its tools, decides when to retry and when to stop. Several 2026 studies set out to measure how much the harness matters relative to the model, holding the model and the tasks fixed, and their answers…
-
Journal article
Repair Without a Repairer: Five Mechanisms Called Self-Repair in Neural Networks, Why Much of the Measured Restoration Is Restored Scale, and What Would License the Claim That a Network Restores Its Own Performance
"Self-repair" is used in at least three research communities that do not cite one another. In transformer interpretability it names the Hydra effect: when one component is ablated, later components shift their output and partly restore the logit. In neuromorphic engineering it names…
-
Journal article
Circuits Are Estimates, Not Mechanisms: A Recovered Subgraph Is One Draw From a Class Set by the Ablation, the Dataset and the Grain, the Results That Defend Circuits Hold at the Level of That Class, and What Ablation Evidence Licenses at Each Level
A circuit in a language model is usually presented as the mechanism by which the model performs a task: a sparse subgraph, found by ablation or patching, that reproduces the behaviour when everything else is knocked out. A second body of work, mostly from 2024 to 2026, reports that the…
-
Journal article
A Control Knob Is Not a Concept: What a Working Steering Vector Licenses, Why Its Equivalence Class Is Set by the Readout That Scores It, and the Controls That Would Upgrade Control to Representation
Adding a vector to a language model's residual stream and watching a behaviour change has become the field's most common argument that the model represents a concept along that vector. Refusal directions, persona vectors, truth directions and instruction vectors are all reported this…
-
Preprint
Théorie du broyeur recycleur et du déséquilibre autogravitatiinnel
L'univers est fini et fermé, un trou noir final broie toute matière comme un broyeur recycleur de métaux, la transforme en énergie pure sous haute pression Pint. Sa gravité G fait le tour de l'univers fini et revient (Gretour ) s'appliquer sur lui même. Gretour +Pint crée une instabilité…
-
Journal article
One Pass, Three Purchases: Deduplication's Efficiency, Generalisation and Privacy Arguments Each Hold in Their Own Regime, Rest on Different Properties of the Repetition Distribution, and Cannot All Be Bought With One Similarity Threshold
Deduplication became a fixed stage of language-model pretraining on the strength of three claims made together: removing duplicates saves compute, improves what the model learns, and reduces memorisation and therefore privacy risk. A reading of the literature from 2024 to 2026 can make…
-
Journal article
Deskilling Is Not One Endpoint: Five Quantities Behind Assisted Gains and Unassisted Losses in Clinical AI, Why Two Studies of the Same Design Disagree, and Why a Within-Clinician Trial's Control Arm Has Already Been Exposed
Clinical AI tools are approved and bought on one endpoint, the performance of a clinician working with the tool, while a second literature asks what happens to the same clinician working without it. The two are usually set against each other as gain against loss. This paper argues that…
-
Journal article
Dictionaries Without Ontologies: What a Sparse Autoencoder Latent Is Evidence For, Why Decodability, Use and Reproducibility Do Not Add Up to Canonicity, and the Identifiability Premise No Real-Model Result Has Checked
Sparse autoencoders are now the default instrument for decomposing language model activations, and their latents are routinely reported as the model's features: units the network itself uses, which a wide enough dictionary has pulled out of superposition. This paper asks what a latent is…
-
Preprint
The Shnoudi Holographic Tensor-Network Construction An Explicit Construction with Independent Computational Cross-Checks on Mathieu-Group Combinatorial Geometry, with a Precisely Scoped Verification of Ryu–Takayanagi Entropy and the Entanglement First Law
This note constructs an explicit tensor network from combinatorial data associated with the Mathieu group M11 and the extended binary Golay code, and tests it against two structures from the holographic tensor-network literature: the Ryu–Takayanagi minimal-cut entropy formula and the…
-
Journal article
Three Hazards Under One Horizon: Constant-Hazard, Degrading-Step and Log-Logistic Accounts of Agent Failure Describe Different Levels of the Same Data, the 50 Percent Trend Barely Depends on Which Is Right, and Every High-Reliability Forecast Does
Forecasts of autonomous AI capability increasingly rest on one number: the length of task, measured in human working time, that an agent completes half the time. That number has doubled roughly every seven months, and three accounts of why agents fail on longer tasks now sit behind its…
-
Journal article
A Contamination Flag Is Not an Inflation Estimate: What Benchmark Contamination Detectors Certify, What Decontamination Is Meant to Buy, and Why the Clean-Subset Comparison That Joins Them Is Biased From Both Sides
Almost every large language model report now includes a contamination analysis: benchmark items that overlap the training corpus are flagged, the model is rescored on the unflagged remainder, and a small gap is read as evidence that the headline numbers stand. Two research literatures…
-
Journal article
No Stable Self: What Self-Referential Anomaly Defence Can Certify for a Continually Learning System, Why Every Located Guarantee Holds the Benign Distribution Still, and Why Immunology's Own Answer Is a Damage Signal From Outside
Immune-inspired security proposes that a system can protect itself by learning what it normally is and treating departures from that as threats. Continual learning makes departure the normal case: the system is supposed to change, so a poisoning campaign and a legitimate shift in the…
-
Journal article
Agreement Is Not a Criterion: What Matching Human Labels Certifies About an LLM Judge, Why a Shared Shortcut Passes the Test, and Why Every External Criterion Found So Far Validates a Use Rather Than a Judge
An automatic judge is usually accepted on one piece of evidence: it agrees with human preference labels about as often as humans agree with each other. That number is then read as permission to use the judge for ranking models, filtering data, selecting samples and supplying reward. This…
-
Journal article
A Stated Limit Is a Forecast: What an Agent Can Truthfully Say About Its Own Permissions, Why the Checkable Part Should Never Come From the Model, and How to Score the Part Only the Model Can Supply
Agent interfaces, skill specifications and disclosure proposals increasingly ask a tool-using language-model agent to tell its user what it may and may not do. The self-explanation literature seems to answer whether such statements can be trusted, and its answer seems to be no. This…
-
Journal article
The Shnoudi Descent: From the Golay Code and Leech Lattice to Four-Dimensional Triality — 24 to 4 Version Two — A Speculative Mathematical Framework for the Emergence of Four-Dimensional Geometry and Particle Symmetry
This document develops a mathematical construction connecting the Golay code and Leech lattice, in twenty-four dimensions, to a four-dimensional, triality-based particle symmetry structure. Every computational claim is verified by 55 independently runnable Python scripts, provided as a…
-
Journal article
Age Is Not Evidence of Staleness: A Recency Decay Is a One-Rate Survival Model, the Rate Belongs to the Fact Rather Than the Clock, and the Silent Change Only a Rate Could Catch Is the Case No Agent-Memory Benchmark Scores
Language agents that keep memories across sessions usually discount a record by how long ago it was written or last read. That discount was introduced to decide what an agent attends to, but it is increasingly read as a statement about whether a record is still true. This paper examines…
-
Journal article
Did the Model Change, or Did Your Measurement? A Drift Claim Is a Test Against a Declared Null, the Careful Longitudinal Studies Mostly Caught Announced Version Moves, and a Capability Change Under a Pinned Identifier Is the Claim the Evidence Supports Least
Two bodies of work describe the same deployed language models in opposite terms. One reports that models behind a fixed product name change over time and calls for continuous monitoring. The other reports that a single model, queried twice under settings meant to be deterministic…
-
Journal article
Detectors Say 'Different', Not 'What': Score, Deviation and Name Are Three Claims With Three Referees, Only the Name Needs a Referee Outside the Detector, and Published Results Show Novelty Detected Far More Reliably Than It Is Named
An anomaly or novelty detector reports that an input is unlike the data it was built from. Users want more: they want to know what the unusual thing is. Explainable anomaly detection is often presented as supplying that second answer, but three different claims travel under the one word…
-
Journal article
The Precursor Assumption: What an Early-Warning Threshold Must Deliver in a Frontier Safety Framework, Why the Continuity Evidence Covers Aggregates Under Fixed Elicitation Rather Than the Thresholded Task, and the Lead Time No Published Record Measures
Frontier safety frameworks decide when a model needs stronger safeguards by testing it against capability thresholds at a fixed cadence, and they place an early-warning threshold below each capability threshold so that mitigations can be prepared in time. That design is sound only if a…
-
Journal article
The Evaluator Is the Bottleneck: What a Self-Modifying System's Acceptance Test Must Satisfy Before Its Verdict Can Be Trusted, Why Co-Evolving Evaluators Shrink the External Anchor Rather Than Remove It, and How Moving the Anchor Outside Moves the Attack Surface With It
Self-modifying AI systems now decide which changes to their own code to keep by running an acceptance test and retaining whatever scores higher. Published coding-agent loops report large benchmark gains this way, and their safety case rests on sandboxes, archives of past versions and the…
-
Journal article
Spend the Budget on the Known or the Unknown? Open-Set Active Learning Scores Two Opposed Objectives Under One Name, Its Filtering Side's Own Results Do Not Treat Unknowns as Waste, and the Winner Is Set by a Relevance Label and a Query Price the Benchmarks Fix in Advance
Active learning chooses which unlabelled examples a human should label. When the unlabelled pool contains classes the model has never been shown, two lines of work give opposite instructions under the same name. One, usually called open-set active learning or active open-set annotation…
-
Dataset
Court Rules: U.S. federal court rules, judge standing orders and filing deadlines dataset
Court Rules is a free reference for U.S. federal court practice: the Federal Rules of Civil Procedure, district local rules, individual judge standing orders and court holidays, organized by court. The deadline calculator counts calendar, business and court days and applies each…
-
Preprint
S H N O U D I R E S E A R C H On the Putative Extremal [72, 36, 16] Binary Code
Whether a binary doubly-even self-dual [72,36,16] code exists has been open since 1973. This document does not settle the question; it maps, with certified exactness, the point at which the known methods stop, so that a future argument can begin there. On the way it proves the results…
-
Journal article
Whose Goals? Autotelic Agents Generate Goals Within Spaces They Are Given: The Goal Space and the Referee Stay Outside the Agent, and a Foundation Model in the Loop Relocates Them Rather Than Removing Them
Autotelic agents are described as learning to represent, generate, select and solve their own goals, and a new generation of systems built on foundation models is reported to do so without hand-coded goal representations, without human intervention, or from zero data. This paper…
-
Journal article
Pooling Experience Spends Independence Twice: Shared Memory in LLM Agent Teams Correlates Errors and Opens a Common-Mode Channel, Why Little of the First Was Left to Spend, and the Order, Gate and Readout That May Decide Whether Pooling Pays
Teams of language-model agents increasingly read and write a common memory: a message pool, a blackboard, a bank of distilled experience, a knowledge base served to several frameworks at once. The case for pooling is efficiency, since no agent rediscovers what another already found. The…
-
Journal article
Problem Choice Without a Referee: An Automated Novelty Check Certifies an Empty Search, a Rediscovery Benchmark Credits the Hypothesis Humans Found Next, and Every Located Referee of Worth Independent of Proposer and Field Needs the Objective Given
Pipelines that claim to automate scientific discovery gate their search on a judgement that a proposed idea or problem is novel and worth pursuing. The best-controlled evidence on that judgement points two ways at once: in a blind study involving more than one hundred NLP researchers…
-
Journal article
Errors That Should Compound, and When They Do: The Product Rule Assumes Independent Steps, Conformal Guarantees Assume Exchangeable Instances, and the Unit of Independence Decides What Each Can Claim
Two bodies of work quantify uncertainty in multi-step language-model reasoning, and both rest on an independence assumption placed at some unit. Step-level error models multiply per-step reliabilities, which, at an illustrative 95% per-step accuracy, predicts that a fifty-step chain…
-
Journal article
Perspectives Without Independence: Multi-Agent and Multi-Persona Reasoning Under Compute-Normalised Comparison, Why the Gains That Survive Are Not the Ones Diversity Predicts, and the Controls That Would Tell Them Apart
Multi-agent debate, multi-persona prompting and related schemes are usually justified by diversity of viewpoint: several perspectives err differently, so their combination is more reliable than any one of them. That justification is a claim about independent errors, and a scheme that…