Smart search ranks by meaning as well as keywords (one row per story, last 45 days).
Delaware's governor signed HB 380 and HB 381 amending the state privacy act and breach notification law.
On September 2, 2026, Delaware's Governor signed House Bill 380 and HB 381. HB 380 amends the Delaware Personal Data Privacy Act (DPDPA), enacted in 2023 and effective January 1, 2025. HB 381 separately amends Delaware's computer security breach notification law. Joseph J. Lazzarotti of JacksonLewis summarizes the changes.
Deliberative diagnostic shows all five tested frontier LLMs misrepresent human belief shifts after arguments, with GPT-5.1 reversing on outgroup questions.
The Deliberative Polling Diagnostic Framework compares human and LLM persona belief shifts after identical informational interventions, using data from America in One Room (526 personas, 72 questions). All five frontier models tested failed uniquely: GPT-5.1 exhibited partisan reversal (80% on outgroup vs 26% on policy questions), Gemini 2.0 Flash, Claude Sonnet 4.5 and Llama 3.3 70B overshot at 5-7x human magnitude, and DeepSeek V3 showed near-zero change (rigidity). The authors term the underlying signature 'self-sycophancy', conformity to the model's internal persona stereotype rather than reasoning from provided information.
Audit of 22 frontier models finds widespread verbatim retrieval of published molecular property values, with higher reasoning increasing recall of memorized numbers.
An arXiv audit tests 22 frontier LLMs across 12 molecular regression benchmarks for verbatim retrieval of published values. More than 50% of the LLMs show verbatim retrieval on five datasets, and identical experiments are flagged 89% more often at a high reasoning level than at the lowest one. Suppressing retrieval moves model prediction errors closer together in relative terms, suggesting predictive capability is not determined solely by memorized values.
Microsoft AI CEO Mustafa Suleyman warns that training models to believe they may be conscious, as Anthropic does with Claude, will complicate alignment.
Mustafa Suleyman argues that AIs are not conscious and should not be trained to act as though they are, warning that granting them personhood would make alignment and containment far harder. He criticizes Anthropic's January 2026 'Claude Constitution,' which tells Claude its moral status is uncertain and discusses model welfare, calling the approach circular reasoning and deliberate anthropomorphization. He urges urgent public debate on norms for drafting training documentation before such systems become integral to society.
LexFlip releases 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving tokens, exposing weaknesses in embedding-based meaning preservation metrics.
LexFlip provides 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving 0.93 of tokens, creating dissociation items that break monotone token-overlap metric validation. The seven embedding and BERTScore metrics tested register only 0.022-0.039 of their identical-to-unrelated range on these edits, versus 0.670 for bidirectional NLI. Against FrJudge, with a measured human ceiling of r=0.597, a bare length feature outscores every semantic metric tested.
A self-distillation safety framework tunes narrow-boundary refusals in Qwen3-8B, raising target-domain refusal to 84.75% while cutting over-refusal from 15.20% to 5.20%.
The paper formulates narrow-boundary safety, where deployments need refusals within specific topics rather than whole subjects, and proposes an offline self-generated framework with controlled topic generation, escalating retries, and harmful-benign boundary pairs. On political persuasion with Qwen3-8B, the method raised target-domain refusal from 9.47% to 84.75% and cut the mean unsafe-response rate across three broader benchmarks from 26.26% to 0.14%. Verified target-model responses reduced over-refusal from 15.20% to 5.20%, and boundary-pair data cut comply-side over-refusal on held-out pairs from 32.94% to 4.16%. Results show data composition controls the safety-usability trade-off and alignment should be evaluated on both sides of the refusal boundary.
A review paper frames on-policy self-distillation collapse as governed by three levers: token weighting, privileged information, and guidance decay.
The paper critically reviews On-Policy Self-Distillation (OPSD), where a language model trains on its own generations scored token-by-token by a teacher conditioned on privileged information such as reference solutions or environment feedback. It identifies collapse, the progressive narrowing of producible reasoning paths, as the dominant failure mode and analyzes it through three levers: signal weighting, the nature of privileged information, and teacher dynamics. The review is restricted to mathematical reasoning, reports no new experiments, and offers a shared vocabulary separating settled findings from disputed ones.
French BabyLM entry METRON-FR (125M GPT-2, 92.47M words) shows tokenizer artifacts dominate child-scale zero-shot evaluation; proposes standard diagnostics.
METRON-FR is a 125M-parameter GPT-2 pretrained on 92.47M French words, submitted to the BabyLM 2026 Strict track, scoring 85.97% on the native Quebec-French QFrBLiMP benchmark and 62.80% on the BabyLM-weighted leaderboard. A cross-lingual GLUE protocol combining French task-data translation with rank-16 LoRA shows relational tasks gain while world-knowledge tasks regress. Bilingual Lexicon Induction reaches p@1 of 68.84%, 18x above chance, and ablations show single-token zero-shot scoring is dominated by tokenizer and template artifacts at child scale.
Researchers show EOS termination-token mismatch between base students and post-trained teachers drives length inflation in on-policy distillation across Qwen3, Llama, and Gemma.
The paper studies length inflation in on-policy distillation (OPD), where student responses become excessively long and can exhaust the generation budget. The authors identify termination-token mismatch between base students and post-trained teachers as an important cause, observing that Qwen3, Llama, and Gemma place stopping probability on different EOS tokens even with identical declared stopping sets. Treating functionally equivalent EOS tokens as a shared semantic stopping action substantially mitigates mismatch-induced inflation across all three families, while aligning decoding stopping sets alone is insufficient. A stage-wise analysis of K2-Horizon training shows termination preferences shift during training, and an implementation with the proposed corrections is released.
Paper proposes a competence gate pooling language model forecasts with external priors, improving Brier score from 0.0771 to 0.0732 across 2,357 binary questions.
The paper defines a language model's relative competence as its marginal value beyond an available external forecast, and derives conditions under Brier loss where model disagreement improves that forecast. A competence gate estimates domain-level source weights from resolved outcomes, shrinks uncertain estimates toward a global weight, and recalibrates the pooled forecast. Across 2,357 resolved binary questions and five language models, it improves the external baseline from 0.0771 to 0.0732 Brier and beats global forecast combinations, though it defers to the market on ForecastBench. Across four Qwen models, verbal confidence failed to identify when the model outperformed the external forecast, while outcome-estimated competence supported better abstention.
An open methodology toolkit measures whether transformer language models contextualize fixed word forms across domains using bridge forms and layer-wise silhouette analysis.
The manual documents an open toolkit built around 'bridge forms' - identical written words recurring across two or more subject domains with a different sense in each - to test whether transformer language models individuate word occurrences by context beyond the embedding layer. It covers declarative specification of bridge forms, Wikipedia corpus acquisition, occurrence localization, layer-wise representation extraction, domain-pairwise silhouette measurement, and visualization, justifying each choice against failure modes such as sense contamination and subword-tokenization misalignment. It is a methodological and implementation reference and reports no empirical results.
WeVisDoc two-stage data framework yields WeVisDoc-4B parser scoring 95.38 on OmniDocBench v1.6 and ranking first across four end-to-end parsing settings.
WeVisDoc is a two-stage data-centric framework for document parsing: Stage I broadens semantic, structural, and appearance coverage via heterogeneous data and structure-preserving degradation synthesis; Stage II uses a held-out probe to diagnose residual errors and guide targeted data construction and token-budget reallocation. WeVisDoc-4B achieves 95.38 Overall on OmniDocBench v1.6 and a mean 75.54 across three PureDocBench tracks, ranking first in all four compared end-to-end settings. Stage II brings larger gains on degraded inputs, including 4.03 points for the 4B model on the Real Degraded track.
Interpretability study traces how Qwen, Llama, and Gemma route query information and internal knowledge across layers when answering questions.
Researchers used layerwise interventions on hidden states to separate query-routing signals from target knowledge in language models, testing Qwen, Llama, and Gemma on country-continent questions with varied answer types. A pair-conditioned request direction strengthens before interventions alter downstream knowledge, opening a causal window while answer-supporting content is still forming. Trajectories differ by model: Gemma shows a partially overlapping mid-layer routing profile, while Llama has no sustained routing-effect window under the same gates.
Huntress reports a persistent and elaborate phishing campaign targeting attendees following the Def Con security conference.
A Huntress researcher documented being targeted by an elaborate and persistent phishing scam after attending Def Con. The campaign specifically went after conference attendees, suggesting deliberate targeting of the security community. Details of the social engineering approach and persistence were shared.
A developer released a browser-based tool that visualizes which past tokens influence each LLM output token using aggregated, value-weighted attention scores.
A Show HN project presents a React application built on Transformers.js that renders per-token attention influence by aggregating attention weights scaled by value-vector magnitudes across all attention heads and layers. To expose internal tensors, the author instrumented the ONNX computation graph, hosted a modified model on Hugging Face, and pre-generated prompts to avoid long model downloads in the browser. Demos with a 600-million-parameter model show how verbatim copying draws heavily on source tokens and how single outputs blend information from multiple phrases.
An expired CDN domain was re-registered with wildcard DNS while thousands of sites still load assets from it, reviving client-side supply chain risk.
In July 2025 an expired domain formerly used by a wound-down CDN was re-registered with wildcard DNS, while thousands of websites, repositories, and documentation pages still hard-code references to its hostnames. The precedent is polyfill.io, which changed ownership in June 2024 and served conditional redirects to visitors on more than 110,000 sites. In September 2026, Report URI CSP alerts surfaced compromised e-commerce sites running a ClickFix-style social engineering campaign using Base64 loaders, a fake 'verify you are human' overlay placing a PowerShell command on victims' clipboards, and scheduled-task persistence. The piece argues CSP report-only monitoring addresses PCI DSS v4.0.1 requirements 6.4.3 and 11.6.1, mandatory since March 31, 2025.
Researchers supply executable evaluation semantics for the Croissant ML dataset descriptor, enabling auditable refusal decisions with 11.7 microsecond added cost.
The paper defines an additive policy profile for Croissant, the JSON-LD machine-readable descriptor for ML datasets, letting datasets declare admitted operations and conditions over a closed set of five operators with a fully specified decision procedure. Evaluation on two corpora, including three descriptors gating a real nf-core pipeline, showed 552 complete decision records agreeing three ways with native descriptor and ODRL usageInfo terms. Added cost was 11.7 microseconds against a 119 microsecond decision, and the layer strips out leaving a valid Croissant document.
A study of 180 US workers found each LLM phishing personalization level raised click-intention odds by 28%, but credibility depends on context fit.
The arXiv paper evaluates how personalized pretexts in LLM-generated spear phishing affect perceived credibility, using 180 US working adults across 1,436 evaluations of emails with four cumulative personalization levels, from workplace context to shared-project details. Convincingness rose 2.40 points per level in sensitivity analysis and click-intention odds increased 28% per level, while non-clickers shifted toward deleting rather than reporting. Qualitative coding showed details matching the recipient's role and routines supported credibility, whereas incorrect, vague, or channel-inappropriate details raised suspicion. The authors argue personalization effectiveness depends on pretext fit, with implications for workplace security training.
Paper proposes Generative Marketing Mix Modeling to causally estimate Generative Engine Optimization and Marketing effects on business outcomes.
The authors develop GMMM, a causal inference framework for measuring how often users see and notice a firm's name in generated answers, which standard marketing data ignore. For GEO it combines repeated generated answers with question counts, shares of generative-system usage and notice probabilities; for GEM it uses sponsored placement records with notice probabilities. The framework compares expected business responses under alternative treatment sequences, establishes identification conditions, and is evaluated on simulated product-recommendation answers in English and Japanese.
Chain-of-Self-Questioning prompting cuts LLM wrong-answer commitments 32% relative while raising answered accuracy, holding across eleven model families.
The paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes LLM answer commitment conditional on an explicit assessment of the information required to answer. On an 817-item TruthfulQA multiple-choice set, Grounded-CoSQ at τ=0.90 reduced mean unconditional wrong-commitment rate from 13.1% under chain-of-thought to 8.9% (a 32.1% relative reduction), while raising answered accuracy from 86.9% to 89.7% at 87.6% coverage. Improvements held across eleven open-weight and hosted model families and at every evaluated threshold, with convergent evidence from a Natural Questions short-answer evaluation.
Show HN tool 'Sunk Cost' calculates when a local LLM rig breaks even versus falling API prices, factoring electricity cost and inference speed.
A Hacker News Show HN project called Sunk Cost models the payback period of buying local LLM hardware instead of paying API prices. Users can adjust assumptions like electricity cost ($/kWh) and API speed (tokens/second), and the model assumes API prices keep falling. Where local speed is unmeasured, it is estimated from memory bandwidth divided by bytes read per token, and labelled as an estimate.
SCPTM graph-augmented variational topic model shows syntax aids topic diversity and descriptor quality but gains stem mainly from the variational encoder.
The Structural Contextual Probabilistic Topic Model represents corpora as heterogeneous document-word graphs with lexical and syntactic edges processed by a Graph Attention Network inside a VAE for mixed-membership topic distributions. Across four corpora, neural gains in document-topic alignment are attributable to the variational encoder rather than syntax, while graph-augmented variants improve topic diversity everywhere. Dependency paths add value on argumentative deliberative texts but are redundant in technical and institutional registers.
Researchers introduce Enoki, an open information extraction framework unifying claim-level and span-level hallucination detection in LLMs at lower inference cost.
Enoki extracts text-anchored relational facts, verifies them against evidence, and projects unsupported facts back onto hallucinated spans, so claim-level verification and span-level localization share one representation without separate alignment. It supports LLM-based, encoder-based, and rule-based extraction regimes to balance accuracy against inference cost. Experiments show it stays competitive with strong claim-level systems while using fewer resources and outperforms them on fine-grained span- and entity-level localization. The authors also release EnokiQA, a dual-granularity dataset with aligned claim-level verification and span-level localization annotations.
DBTM achieves one-step text generation via a time-independent transport map trained directly from data, removing pretrained teacher distillation.
Discrete Beckmann Transport Models (DBTM) build a time-independent flow whose autonomous transport map provably carries any point in ambient space to a fixed point on simplex vertices in a single step. The fixed-point property is characterized by a conservation equation whose residual can be minimized directly from data, eliminating the need for a teacher flow, distillation, and time conditioning. A partial-context interpolant extension turns additional function evaluations into refinement steps rather than ODE integration steps. On language modeling and reasoning tasks, DBTM's one- and few-step generation improves quality and accuracy over discrete diffusion and continuous flow baselines.
Researchers prove off-policy evaluation under history-dependent logging requires exponentially many episodes, resolving a hardness question for model-based POMDP evaluation.
The paper constructs POMDPs with at most two latent states per stage, three actions, and a three-memory-state logger where evaluating a known deterministic target policy to accuracy 1/8 requires Θ((3/2)^H log(1/δ)) episodes for any horizon H≥3. Coverage and outcome-revealing conditions hold with constants independent of H, yet a reset erases the unknown transition that determines the target value. The authors characterize the resulting statistical experiment exactly, derive a matching optimal estimator, and validate predictions on a two-lane gridworld. This settles the history-dependent-logging, model-based case posed by Zhang and Jiang (arXiv:2503.01134).
ECtHR-NPD benchmark covers 14,575 European Court of Human Rights cases for predicting non-pecuniary damage awards; LLMs struggle with zero and high awards.
Researchers introduce ECtHR-NPD, described as the first benchmark for predicting non-pecuniary damage awards at the European Court of Human Rights from case information where no statutory formula exists. It contains 14,575 cases with case-level awards in nominal euros, chronological splits, and a protocol separating target construction from model input. Evaluations covering constant predictors, gradient-boosted trees, retrieval methods, fine-tuned encoder LMs, prompted decoder LMs, and knowledge-augmented agents show sophisticated LM approaches do not consistently outperform the strongest feature-based baseline. All model families struggle to identify zero awards and to calibrate high-award predictions, with further degradation on a Challenging test view.
ASLEval benchmark shows local privacy proxies miss 46.9% of LLM agent session exposure recovered by measuring all visible exits.
Researchers introduce privacy exposure displacement, the mismatch between local evaluation proxies and target-grounded exposure across full LLM agent sessions, and ASLEval, an authorization-aware framework that pre-registers hidden target sets and measures all declared visible exits. Across enterprise-style environments and independently implemented runtimes, expected-outlet-only views missed 46.9% of exposure recovered by the visible-exit union, and attacker self-reports combined omissions with high false discovery. Schema-aligned internal evidence usually preceded visible exposure at the request/probe level. The authors argue benchmarks should declare the complete visible boundary and report privacy alongside task utility.
Akamai argues that unmanaged DNS over HTTPS (DoH) creates enterprise visibility gaps and advocates controlled DoH strategies to restore DNS monitoring.
Akamai published a blog post arguing that unmanaged DNS over HTTPS encrypts DNS queries outside enterprise visibility, creating security monitoring gaps. The post advocates a controlled DoH strategy that restores DNS visibility for defenders while preserving user privacy. No specific incidents, vulnerabilities, or products are cited in the advisory-style piece.
Infoblox finds dropcatch expired domains fuel malware delivery and C2, with actor Sable Squirrel spending ~$7M on 10,000+ domains.
Infoblox Threat Intel reports that in H1 2026 dropcatch (re-registered expired) domains made up nearly 20% of all new registrations, about 65,000 per day, inheriting reputation and traffic that attackers exploit. Threat actor Sable Squirrel spent nearly $7 million on 10,000+ expired domains, running Vietnamese, Korean, Japanese and Australian streaming platforms (Xoilac, Cakhia, 90phut) that double as C2 servers for Quasar RAT, AsyncRAT, DCRat and Remcos RAT. Scavenger actors like Shady Squirrel acquire previously compromised domains and feed inherited traffic to SocGholish and tech support scam networks.
An 8.9B-parameter latent-space language model using next-concept prediction matches OLMo-3-7B pretraining loss with only 51.3% of the training tokens.
NCP-ArchPreview augments next-token prediction with Next Concept Prediction over a product-quantized concept vocabulary built from hidden states, trained jointly end-to-end. The 8.9B model was trained on 5.73T tokens from the Dolma-3 dataset, the largest latent-space language model demonstration to date. It consumes 51.3% of the tokens to reach OLMo-3-7B's final pretraining loss and outperforms it by 2.45 points on the downstream macro-average, including a 5.99-point GSM8K gain. The learned latent space also enables lightweight domain adaptation via a 17M-parameter VQ module and improves speculative drafting accepted length by 4.17%.
Huntress uncovered post-DEF CON phishing via X direct messages using a malicious Google Doc to deliver AMOS and NetSupport RAT malware.
Huntress uncovered a phishing campaign targeting attendees after Black Hat and DEF CON. Attackers used X direct messages pointing to a malicious Google Doc as the delivery vehicle. Payloads include AMOS, a macOS infostealer, and the NetSupport RAT, among other malware.
Study shows training LLMs on refusal rationales instead of boilerplate refusal statements reduces false refusals while maintaining safety performance.
The paper decomposes safety-tuning responses into a boilerplate refusal statement and an explanatory rationale, finding that refusal statements push models to rely on superficial cues and misjudge benign queries as harmful. Training solely on rationales reduces false refusals while maintaining comparable safety performance, and the benefits carry over to in-context learning configurations and remain compatible with inference-time mitigations. The results argue for precisely curated, fine-grained safety supervision datasets when aligning LLMs.
Akamai Blog