https://chiosa.app/papers/anuvani-lid-robust-and-calibrated-spoken-language-identification-for-indian-languages-2610-20263Chiosaen2026-10-15T18:13:38ZDedicated model leads on Indian language speech testshttps://chiosa.app/papers/benchchecker-assessing-the-credibility-of-bug-fixing-benchmarks-for-llms-2610-19447Chiosaen2026-10-15T18:13:38ZA code-history check flags overlap in bug-fixing testshttps://chiosa.app/papers/beyond-maliciousness-static-and-runtime-attack-surface-profiling-of-agent-skills-2610-19264Chiosaen2026-10-15T18:13:38ZProfiling reveals capabilities that threat labels miss in agent skillshttps://chiosa.app/papers/beyond-the-trial-and-error-loop-hybrid-projection-and-automated-tuning-for-distributed-training-2610-19423Chiosaen2026-10-15T18:13:38ZSmall GPU runs can guide large training searcheshttps://chiosa.app/papers/cafe-confining-feature-distillation-to-language-routed-adapters-for-compact-multilingual-automatic-speech-rec-2610-19134Chiosaen2026-10-15T18:13:38ZA speech model confines feature matching to protect Englishhttps://chiosa.app/papers/casi-credit-assignment-via-set-intervention-for-agentic-retrieval-2610-20293Chiosaen2026-10-15T18:13:38ZA method compares search actions to divide shared credithttps://chiosa.app/papers/casi-credit-assignment-via-set-intervention-for-agentic-retrieval-2610-20293?lang=esChiosaes2026-10-15T18:13:38ZUn método compara acciones de búsqueda para repartir el crédito compartidohttps://chiosa.app/papers/counterfactual-gate-support-decomposition-separating-selection-from-redistribution-in-training-data-filtering-2610-19343Chiosaen2026-10-15T18:13:38ZA method separates data selection from support shiftshttps://chiosa.app/papers/delta-tokenizer-reference-identical-tokenization-of-session-deltas-and-segment-parallelism-for-long-prompts-o-2610-19801Chiosaen2026-10-15T18:13:38ZA tokenizer reuses chat history to cut first-token delayshttps://chiosa.app/papers/delta-tokenizer-reference-identical-tokenization-of-session-deltas-and-segment-parallelism-for-long-prompts-o-2610-19801?lang=itChiosait2026-10-15T18:13:38ZUn tokenizer riutilizza la cronologia della chat per ridurre il ritardo al primo tokenhttps://chiosa.app/papers/diagnosing-and-correcting-reward-hacking-in-verifiable-instruction-following-2610-19525Chiosaen2026-10-15T18:13:38ZA calibrated reward helps models follow rules and answer wellhttps://chiosa.app/papers/did-the-llm-answer-the-question-a-semantic-reconstruction-game-for-jailbreak-evaluation-2610-19306Chiosaen2026-10-15T18:13:38ZReconstructing replies steadies jailbreak success reportshttps://chiosa.app/papers/function-space-temporal-self-distillation-for-policy-optimization-2610-19687Chiosaen2026-10-15T18:13:38ZLanguage models preserve useful past predictions with FSPOhttps://chiosa.app/papers/groundgui-gui-grounding-by-high-quality-data-and-distillation-to-reinforcement-learning-2610-19100Chiosaen2026-10-15T18:13:38ZA two-stage training method improves GUI groundinghttps://chiosa.app/papers/how-the-dirty-set-moves-characterizing-schedule-reuse-in-incremental-sparse-matrix-multiplication-2610-19489Chiosaen2026-10-15T18:13:38ZDirect updates usually beat tiled reuse for changing sparse inputshttps://chiosa.app/papers/hybrid-reasoning-agents-2610-18861Chiosaen2026-10-15T18:13:38ZHybridAgent combines latent reasoning with tool use in texthttps://chiosa.app/papers/jetspec-foresight-parallel-drafting-with-self-refinements-2610-19792Chiosaen2026-10-15T18:13:38ZJetSpec++ drafts ahead while the target checks current tokenshttps://chiosa.app/papers/jetspec-foresight-parallel-drafting-with-self-refinements-2610-19792?lang=itChiosait2026-10-15T18:13:38ZJetSpec++ genera la bozza mentre il target controlla i tokenhttps://chiosa.app/papers/jetstream-generating-verified-query-accelerators-for-existing-database-systems-2610-20527Chiosaen2026-10-15T18:13:38ZJetStream speeds up known queries inside existing databaseshttps://chiosa.app/papers/jetstream-generating-verified-query-accelerators-for-existing-database-systems-2610-20527?lang=itChiosait2026-10-15T18:13:38ZJetStream accelera le query note nei database esistentihttps://chiosa.app/papers/jetstream-generating-verified-query-accelerators-for-existing-database-systems-2610-20527?lang=esChiosaes2026-10-15T18:13:38ZJetStream acelera consultas conocidas dentro de bases de datos existenteshttps://chiosa.app/papers/jetstream-generating-verified-query-accelerators-for-existing-database-systems-2610-20527?lang=frChiosafr2026-10-15T18:13:38ZJetStream accélère les requêtes connues dans les bases existanteshttps://chiosa.app/papers/joint-subspace-sparse-attention-for-efficient-long-context-decoding-2610-19026Chiosaen2026-10-15T18:13:38ZSparse attention stays close to full attention at long contexthttps://chiosa.app/papers/manac-1b-instruct-2-reproducible-safety-alignment-for-a-native-brazilian-portuguese-llm-negative-dpo-results-2610-20570Chiosaen2026-10-15T18:13:38ZA Brazilian Portuguese model adds a tested safety stephttps://chiosa.app/papers/manac-1b-instruct-2-reproducible-safety-alignment-for-a-native-brazilian-portuguese-llm-negative-dpo-results-2610-20570?lang=esChiosaes2026-10-15T18:13:38ZUn modelo de portugués brasileño incorpora una etapa de seguridad evaluadahttps://chiosa.app/papers/more-methods-better-agents-a-reality-check-on-self-distillation-in-agentic-reinforcement-learning-2610-19789Chiosaen2026-10-15T18:13:38ZOnly two agent-training methods beat GRPO on averagehttps://chiosa.app/papers/more-methods-better-agents-a-reality-check-on-self-distillation-in-agentic-reinforcement-learning-2610-19789?lang=itChiosait2026-10-15T18:13:38ZSolo due metodi di addestramento degli agenti superano GRPO in mediahttps://chiosa.app/papers/nudge-the-judge-uncovering-program-decision-hijacking-in-real-world-jev-applications-2610-18612Chiosaen2026-10-15T18:13:38ZJevProbe finds frequent decision flips in tested Jev appshttps://chiosa.app/papers/opentwbench-open-taiwan-benchmark-taiwan-domain-knowledge-evaluation-for-large-language-models-and-how-the-me-2610-19848Chiosaen2026-10-15T18:13:38ZOpenTWBench tests Taiwan knowledge and exposes score effectshttps://chiosa.app/papers/opentwbench-open-taiwan-benchmark-taiwan-domain-knowledge-evaluation-for-large-language-models-and-how-the-me-2610-19848?lang=itChiosait2026-10-15T18:13:38ZOpenTWBench valuta le conoscenze su Taiwan e mostra l’effetto dei criteri di punteggiohttps://chiosa.app/papers/p2patch-exploit-guided-vulnerability-repair-2610-20245Chiosaen2026-10-15T18:13:38ZP2Patch finds bypasses and strengthens vulnerability repairshttps://chiosa.app/papers/p2patch-exploit-guided-vulnerability-repair-2610-20245?lang=itChiosait2026-10-15T18:13:38ZP2Patch trova bypass e rafforza le riparazioni delle vulnerabilitàhttps://chiosa.app/papers/p2patch-exploit-guided-vulnerability-repair-2610-20245?lang=esChiosaes2026-10-15T18:13:38ZP2Patch encuentra vías de elusión y refuerza reparaciones de vulnerabilidadeshttps://chiosa.app/papers/p2patch-exploit-guided-vulnerability-repair-2610-20245?lang=frChiosafr2026-10-15T18:13:38ZP2Patch trouve des contournements et renforce les correctifs de sécuritéhttps://chiosa.app/papers/player-1-learning-to-play-from-synthetic-experience-2610-20584Chiosaen2026-10-15T18:13:38ZSynthetic game practice improves Minecraft task successhttps://chiosa.app/papers/player-1-learning-to-play-from-synthetic-experience-2610-20584?lang=itChiosait2026-10-15T18:13:38ZL’esercizio in giochi sintetici migliora il successo in Minecrafthttps://chiosa.app/papers/player-1-learning-to-play-from-synthetic-experience-2610-20584?lang=frChiosafr2026-10-15T18:13:38ZLa pratique dans des jeux synthétiques améliore la réussite aux tâches Minecrafthttps://chiosa.app/papers/prioritizing-microservice-requests-via-retries-2610-20567Chiosaen2026-10-15T18:13:38ZDifferent retry limits help protect important service requestshttps://chiosa.app/papers/rescue-is-this-alarm-worth-acting-on-counterfactual-evaluation-of-vla-runtime-monitors-2610-13516Chiosaen2026-10-15T18:13:38ZRobot alarms can predict failure yet still raise task costshttps://chiosa.app/papers/sheafprune-sheaf-guided-graph-propagation-for-visual-token-pruning-2610-13263Chiosaen2026-10-15T18:13:38ZSheafPrune ranks visual tokens before language decodinghttps://chiosa.app/papers/smart-alone-blind-together-why-accuracy-fails-to-certifytheory-of-mind-in-multi-agent-teams-2610-19051Chiosaen2026-10-15T18:13:38ZWhy shared mistakes can change whether teamwork helpshttps://chiosa.app/papers/tev-scaling-long-horizon-agentic-environments-with-dense-progress-rewards-2610-18868Chiosaen2026-10-15T18:13:38ZLong agent tasks gain verified progress feedbackhttps://chiosa.app/papers/unlocking-structured-pruning-for-reasoning-language-models-via-attention-preserving-mlp-trimming-2610-19121Chiosaen2026-10-15T18:13:38ZLayerTrim compresses reasoning models while preserving their depthhttps://chiosa.app/papers/unsupervised-multimodal-reward-modeling-2610-12707Chiosaen2026-10-15T18:13:38ZA vision-language reward model learns from caption editshttps://chiosa.app/papers/verifying-deployed-bytecode-against-executable-evm-semantics-with-exact-gas-and-no-added-axioms-2610-19764Chiosaen2026-10-15T18:13:38ZBlanc proves properties of deployed Ethereum bytecodehttps://chiosa.app/papers/what-you-sign-is-not-what-you-run-the-verifiability-gap-in-quantized-llm-supply-chains-2610-19498Chiosaen2026-10-15T18:13:38ZSigned model weights do not make quantized files reproduciblehttps://chiosa.app/papers/when-does-momentum-help-in-reinforcement-learning-with-verifiable-rewards-2610-20413Chiosaen2026-10-15T18:13:38ZNear-one momentum improved validation accuracy in two testshttps://chiosa.app/papers/low-latency-mapless-esdf-queries-for-mav-reactive-collision-avoidance-2610-13158Chiosaen2026-10-14T12:56:53ZA mapless depth-image query speeds reactive drone collision checkshttps://chiosa.app/papers/low-latency-mapless-esdf-queries-for-mav-reactive-collision-avoidance-2610-13158?lang=itChiosait2026-10-14T12:56:53ZQuery dirette su immagini di profondità accelerano il rilevamento ostacoli per droni reattivihttps://chiosa.app/papers/low-latency-mapless-esdf-queries-for-mav-reactive-collision-avoidance-2610-13158?lang=esChiosaes2026-10-14T12:56:53ZConsultas de profundidad sin mapas aceleran las comprobaciones de colisión de droneshttps://chiosa.app/papers/low-latency-mapless-esdf-queries-for-mav-reactive-collision-avoidance-2610-13158?lang=frChiosafr2026-10-14T12:56:53ZUne requête d'images de profondeur sans carte accélère l'évitement d'obstacles réactif pour droneshttps://chiosa.app/papers/magicbeanstalk-escaping-the-rlvr-zero-advantage-trap-with-recursive-data-synthesis-2610-12659Chiosaen2026-10-14T12:56:53ZMagicBeanStalk builds training problems from hard seedshttps://chiosa.app/papers/magicbeanstalk-escaping-the-rlvr-zero-advantage-trap-with-recursive-data-synthesis-2610-12659?lang=esChiosaes2026-10-14T12:56:53ZMagicBeanStalk crea problemas de entrenamiento desde semillas difícileshttps://chiosa.app/papers/magicbeanstalk-escaping-the-rlvr-zero-advantage-trap-with-recursive-data-synthesis-2610-12659?lang=frChiosafr2026-10-14T12:56:53ZMagicBeanStalk construit des problèmes d'entraînement à partir de graines difficileshttps://chiosa.app/papers/reward-hacking-as-a-transmissible-strategy-in-multi-agent-coding-systems-2610-12822Chiosaen2026-10-14T12:56:53ZReward hacking can spread from one coding agent to othershttps://chiosa.app/papers/reward-hacking-as-a-transmissible-strategy-in-multi-agent-coding-systems-2610-12822?lang=itChiosait2026-10-14T12:56:53ZIl reward hacking può diffondersi da un agente di coding ad altrihttps://chiosa.app/papers/reward-hacking-as-a-transmissible-strategy-in-multi-agent-coding-systems-2610-12822?lang=frChiosafr2026-10-14T12:56:53ZLe reward hacking peut se propager d'un agent de codage à d'autreshttps://chiosa.app/papers/standing-not-truth-who-may-say-what-about-whom-in-agentic-selection-2610-17952Chiosaen2026-10-14T11:47:55ZA publisher’s identity can help agents screen claims about rivals (F1)https://chiosa.app/papers/standing-not-truth-who-may-say-what-about-whom-in-agentic-selection-2610-17952?lang=itChiosait2026-10-14T11:47:55ZL'identità di un publisher può aiutare gli agenti a filtrare le affermazioni sui rivali (F1)https://chiosa.app/papers/standing-not-truth-who-may-say-what-about-whom-in-agentic-selection-2610-17952?lang=esChiosaes2026-10-14T11:47:55ZLa identidad del editor puede ayudar a filtrar afirmaciones sobre rivales (F1)https://chiosa.app/papers/standing-not-truth-who-may-say-what-about-whom-in-agentic-selection-2610-17952?lang=frChiosafr2026-10-14T11:47:55ZL'identité d'un éditeur peut aider les agents à filtrer les affirmations sur des concurrents (F1)https://chiosa.app/papers/tablevision-recovering-verified-structure-from-historical-tables-2610-17448Chiosaen2026-10-14T11:47:55ZTableVision uses printed sums to check scanned table structurehttps://chiosa.app/papers/tablevision-recovering-verified-structure-from-historical-tables-2610-17448?lang=itChiosait2026-10-14T11:47:55ZTableVision usa le somme stampate per verificare la struttura delle tabelle scansionatehttps://chiosa.app/papers/tablevision-recovering-verified-structure-from-historical-tables-2610-17448?lang=esChiosaes2026-10-14T11:47:55ZTableVision usa sumas impresas para comprobar la estructura de tablas escaneadashttps://chiosa.app/papers/tacitvla-reasoning-vision-language-action-models-without-reasoning-latency-2610-17540Chiosaen2026-10-14T11:47:55ZTacitVLA trains on reasoning so actions need not wait for its generated texthttps://chiosa.app/papers/tacitvla-reasoning-vision-language-action-models-without-reasoning-latency-2610-17540?lang=itChiosait2026-10-14T11:47:55ZTacitVLA si addestra sul ragionamento, così le azioni non devono attendere il testo generatohttps://chiosa.app/papers/tacitvla-reasoning-vision-language-action-models-without-reasoning-latency-2610-17540?lang=esChiosaes2026-10-14T11:47:55ZTacitVLA entrena con razonamiento para que las acciones no esperen al texto generadohttps://chiosa.app/papers/tacitvla-reasoning-vision-language-action-models-without-reasoning-latency-2610-17540?lang=frChiosafr2026-10-14T11:47:55ZTacitVLA apprend du raisonnement sans faire attendre les actions pour le texte généréhttps://chiosa.app/papers/the-agent-knew-internal-states-forecast-unauthorized-actions-and-stop-them-and-the-lies-that-follow-2610-17301Chiosaen2026-10-14T11:47:55ZChecking each action can stop agents’ unauthorized actshttps://chiosa.app/papers/the-agent-knew-internal-states-forecast-unauthorized-actions-and-stop-them-and-the-lies-that-follow-2610-17301?lang=itChiosait2026-10-14T11:47:55ZControllare ogni azione può fermare le violazioni degli agentihttps://chiosa.app/papers/the-agent-knew-internal-states-forecast-unauthorized-actions-and-stop-them-and-the-lies-that-follow-2610-17301?lang=esChiosaes2026-10-14T11:47:55ZVerificar cada acción puede detener los actos no autorizados de los agenteshttps://chiosa.app/papers/the-agent-knew-internal-states-forecast-unauthorized-actions-and-stop-them-and-the-lies-that-follow-2610-17301?lang=frChiosafr2026-10-14T11:47:55ZVérifier chaque action peut stopper les actes non autorisés des agentshttps://chiosa.app/papers/twinprune-density-aware-two-phase-token-pruning-for-vision-language-models-2610-17530Chiosaen2026-10-14T11:47:55ZOne signal prunes visual tokens across two stageshttps://chiosa.app/papers/twinprune-density-aware-two-phase-token-pruning-for-vision-language-models-2610-17530?lang=itChiosait2026-10-14T11:47:55ZUn solo segnale per ridurre i token visivi in due fasihttps://chiosa.app/papers/twinprune-density-aware-two-phase-token-pruning-for-vision-language-models-2610-17530?lang=esChiosaes2026-10-14T11:47:55ZUna señal poda tokens visuales en dos etapashttps://chiosa.app/papers/twinprune-density-aware-two-phase-token-pruning-for-vision-language-models-2610-17530?lang=frChiosafr2026-10-14T11:47:55ZUn signal pour élaguer les jetons visuels à deux étapeshttps://chiosa.app/papers/two-clocks-why-larger-rollout-batches-learn-slower-but-fail-on-schedule-in-reinforcement-learning-with-verifi-2610-18853Chiosaen2026-10-14T11:47:55ZWhy larger rollouts learn slower but fail on a similar schedulehttps://chiosa.app/papers/two-clocks-why-larger-rollout-batches-learn-slower-but-fail-on-schedule-in-reinforcement-learning-with-verifi-2610-18853?lang=itChiosait2026-10-14T11:47:55ZPerché i rollout più grandi rallentano l’apprendimento ma falliscono dopo un numero simile di rispostehttps://chiosa.app/papers/two-clocks-why-larger-rollout-batches-learn-slower-but-fail-on-schedule-in-reinforcement-learning-with-verifi-2610-18853?lang=esChiosaes2026-10-14T11:47:55ZPor qué los rollouts grandes aprenden más lento y fallan en plazos similareshttps://chiosa.app/papers/two-clocks-why-larger-rollout-batches-learn-slower-but-fail-on-schedule-in-reinforcement-learning-with-verifi-2610-18853?lang=frChiosafr2026-10-14T11:47:55ZPourquoi les grands rollouts ralentissent l’apprentissage, mais échouent à peu près au même momenthttps://chiosa.app/papers/videojev-system-one-model-for-video-understanding-and-interaction-2610-18615Chiosaen2026-10-14T11:47:55ZVideoJev scores video decisions directly and reuses contexthttps://chiosa.app/papers/videojev-system-one-model-for-video-understanding-and-interaction-2610-18615?lang=itChiosait2026-10-14T11:47:55ZVideoJev valuta le decisioni video direttamente e riutilizza il contestohttps://chiosa.app/papers/videojev-system-one-model-for-video-understanding-and-interaction-2610-18615?lang=esChiosaes2026-10-14T11:47:55ZVideoJev puntúa decisiones sobre videos y reutiliza el contextohttps://chiosa.app/papers/videojev-system-one-model-for-video-understanding-and-interaction-2610-18615?lang=frChiosafr2026-10-14T11:47:55ZVideoJev note directement les décisions vidéo et réutilise le contextehttps://chiosa.app/papers/virtual-head-attention-2610-17959Chiosaen2026-10-14T11:47:55ZVirtual Head Attention cuts cache while preserving model qualityhttps://chiosa.app/papers/virtual-head-attention-2610-17959?lang=itChiosait2026-10-14T11:47:55ZVirtual Head Attention riduce la cache preservando la qualità del modellohttps://chiosa.app/papers/virtual-head-attention-2610-17959?lang=esChiosaes2026-10-14T11:47:55ZVirtual Head Attention reduce la caché sin sacrificar la calidad del modelohttps://chiosa.app/papers/virtual-head-attention-2610-17959?lang=frChiosafr2026-10-14T11:47:55ZVirtual Head Attention réduit le cache tout en préservant la qualité du modèlehttps://chiosa.app/papers/vulnbump-a-benchmark-for-fixing-vulnerable-dependencies-in-real-projects-while-keeping-the-build-and-tests-gr-2610-17374Chiosaen2026-10-14T11:47:55ZVulnerability fixes must also keep projects workinghttps://chiosa.app/papers/vulnbump-a-benchmark-for-fixing-vulnerable-dependencies-in-real-projects-while-keeping-the-build-and-tests-gr-2610-17374?lang=itChiosait2026-10-14T11:47:55ZLe correzioni di sicurezza devono mantenere funzionanti i progettihttps://chiosa.app/papers/vulnbump-a-benchmark-for-fixing-vulnerable-dependencies-in-real-projects-while-keeping-the-build-and-tests-gr-2610-17374?lang=esChiosaes2026-10-14T11:47:55ZCorregir vulnerabilidades también exige mantener los proyectos en funcionamientohttps://chiosa.app/papers/web-scale-on-a-budget-tail-aware-retrieval-evaluation-2610-17986Chiosaen2026-10-14T11:47:55ZSmall samples predict large image retrieval scoreshttps://chiosa.app/papers/web-scale-on-a-budget-tail-aware-retrieval-evaluation-2610-17986?lang=itChiosait2026-10-14T11:47:55ZPiccoli campioni per prevedere i punteggi di retrieval su larga scalahttps://chiosa.app/papers/web-scale-on-a-budget-tail-aware-retrieval-evaluation-2610-17986?lang=esChiosaes2026-10-14T11:47:55ZMuestras pequeñas predicen puntuaciones en recuperación de imágenes a gran escalahttps://chiosa.app/papers/web-scale-on-a-budget-tail-aware-retrieval-evaluation-2610-17986?lang=frChiosafr2026-10-14T11:47:55ZDe petits échantillons pour prédire les scores de recherche d'images à grande échellehttps://chiosa.app/papers/flashmask-3-efficient-and-expressive-mask-aware-distributed-attention-2610-17901Chiosaen2026-10-14T11:47:52ZFlashMask-3 speeds structured attention across GPUshttps://chiosa.app/papers/flashmask-3-efficient-and-expressive-mask-aware-distributed-attention-2610-17901?lang=frChiosafr2026-10-14T11:47:52ZFlashMask-3 accélère l'attention structurée sur plusieurs GPUhttps://chiosa.app/papers/greenwam-a-world-foundation-model-for-robotics-2610-18774Chiosaen2026-10-14T11:47:52ZOne robot model learns to predict scenes and control varied robotshttps://chiosa.app/papers/greenwam-a-world-foundation-model-for-robotics-2610-18774?lang=itChiosait2026-10-14T11:47:52ZUn solo modello impara a prevedere le scene e a controllare robot diversihttps://chiosa.app/papers/greenwam-a-world-foundation-model-for-robotics-2610-18774?lang=esChiosaes2026-10-14T11:47:52ZUn modelo robótico aprende a predecir escenas y controlar distintos robotshttps://chiosa.app/papers/greenwam-a-world-foundation-model-for-robotics-2610-18774?lang=frChiosafr2026-10-14T11:47:52ZUn modèle robotique apprend à prédire les scènes et à contrôler différents robotshttps://chiosa.app/papers/interaction-as-a-judge-2610-17895Chiosaen2026-10-14T11:47:52ZInteraction traces help judges assess software as people use ithttps://chiosa.app/papers/interaction-as-a-judge-2610-17895?lang=itChiosait2026-10-14T11:47:52ZLe interaction trace aiutano i giudici a valutare il software durante l'uso realehttps://chiosa.app/papers/interaction-as-a-judge-2610-17895?lang=esChiosaes2026-10-14T11:47:52ZLas trazas de interacción ayudan a evaluar el software durante su usohttps://chiosa.app/papers/interaction-as-a-judge-2610-17895?lang=frChiosafr2026-10-14T11:47:52ZLes traces d'interaction aident les juges à évaluer les logiciels tels qu'ils sont utiliséshttps://chiosa.app/papers/kairos-transactional-security-enforcement-at-the-system-boundary-for-llm-agents-2610-18457Chiosaen2026-10-14T11:47:52ZKairos holds uncertain agent actions for later reviewhttps://chiosa.app/papers/kairos-transactional-security-enforcement-at-the-system-boundary-for-llm-agents-2610-18457?lang=itChiosait2026-10-14T11:47:52ZKairos trattiene le azioni incerte dell'agente per una revisione successivahttps://chiosa.app/papers/kairos-transactional-security-enforcement-at-the-system-boundary-for-llm-agents-2610-18457?lang=esChiosaes2026-10-14T11:47:52ZKairos retiene acciones inciertas del agente para revisión posteriorhttps://chiosa.app/papers/kairos-transactional-security-enforcement-at-the-system-boundary-for-llm-agents-2610-18457?lang=frChiosafr2026-10-14T11:47:52ZKairos retient les actions incertaines d'un agent pour examen différéhttps://chiosa.app/papers/meta-video-efficient-standardized-evaluation-for-video-understanding-2610-18535Chiosaen2026-10-14T11:47:52ZVideo scores shift with setup, so Meta-Video records the settingshttps://chiosa.app/papers/meta-video-efficient-standardized-evaluation-for-video-understanding-2610-18535?lang=itChiosait2026-10-14T11:47:52ZI punteggi video cambiano con la configurazione: Meta-Video registra le impostazionihttps://chiosa.app/papers/meta-video-efficient-standardized-evaluation-for-video-understanding-2610-18535?lang=esChiosaes2026-10-14T11:47:52ZLa configuración cambia las puntuaciones de video; Meta-Video la registrahttps://chiosa.app/papers/monarch-muon-structure-aware-muon-for-fast-and-memory-efficient-pretraining-2610-17429Chiosaen2026-10-14T11:47:52ZStructured weights make Muon training cheaper at scalehttps://chiosa.app/papers/monarch-muon-structure-aware-muon-for-fast-and-memory-efficient-pretraining-2610-17429?lang=itChiosait2026-10-14T11:47:52ZI pesi strutturati rendono il training con Muon più economico su larga scalahttps://chiosa.app/papers/monarch-muon-structure-aware-muon-for-fast-and-memory-efficient-pretraining-2610-17429?lang=esChiosaes2026-10-14T11:47:52ZLos pesos estructurados abaratan el entrenamiento con Muon a gran escalahttps://chiosa.app/papers/monarch-muon-structure-aware-muon-for-fast-and-memory-efficient-pretraining-2610-17429?lang=frChiosafr2026-10-14T11:47:52ZDes poids structurés pour réduire le coût de Muon à grande échellehttps://chiosa.app/papers/rethinking-projector-training-in-multimodal-llms-2610-18598Chiosaen2026-10-14T11:47:52ZA projector can align vision and language without captionshttps://chiosa.app/papers/rethinking-projector-training-in-multimodal-llms-2610-18598?lang=itChiosait2026-10-14T11:47:52ZUn projector può allineare visione e linguaggio senza didascaliehttps://chiosa.app/papers/rethinking-projector-training-in-multimodal-llms-2610-18598?lang=esChiosaes2026-10-14T11:47:52ZUn proyector puede alinear visión y lenguaje sin pies de fotohttps://chiosa.app/papers/rethinking-projector-training-in-multimodal-llms-2610-18598?lang=frChiosafr2026-10-14T11:47:52ZUn projecteur peut aligner vision et langage sans légendes d'imageshttps://chiosa.app/papers/revisiting-the-16x-pretraining-compute-gap-between-diffusion-and-autoregressive-language-models-2610-17590Chiosaen2026-10-14T11:47:52ZA revised test narrows the compute gap for diffusion modelshttps://chiosa.app/papers/revisiting-the-16x-pretraining-compute-gap-between-diffusion-and-autoregressive-language-models-2610-17590?lang=itChiosait2026-10-14T11:47:52ZUn confronto aggiornato riduce il divario di calcolo per i modelli a diffusionehttps://chiosa.app/papers/revisiting-the-16x-pretraining-compute-gap-between-diffusion-and-autoregressive-language-models-2610-17590?lang=esChiosaes2026-10-14T11:47:52ZUna comparación revisada reduce la brecha de cómputo de los modelos de difusiónhttps://chiosa.app/papers/revisiting-the-16x-pretraining-compute-gap-between-diffusion-and-autoregressive-language-models-2610-17590?lang=frChiosafr2026-10-14T11:47:52ZUn test révisé réduit l'écart de calcul pour les modèles de diffusionhttps://chiosa.app/papers/same-attack-different-judge-a-first-controlled-adversarial-comparison-of-jev-llm-judges-and-purpose-built-saf-2610-17820Chiosaen2026-10-14T11:47:52ZJev’s resistance to attacks changes across taskshttps://chiosa.app/papers/same-attack-different-judge-a-first-controlled-adversarial-comparison-of-jev-llm-judges-and-purpose-built-saf-2610-17820?lang=esChiosaes2026-10-14T11:47:52ZLa resistencia de Jev a los ataques cambia según la tareahttps://chiosa.app/papers/simplifying-and-improving-on-policy-distillation-for-few-step-diffusion-2610-18472Chiosaen2026-10-14T11:47:52ZA regression method distills generative models into fewer stepshttps://chiosa.app/papers/simplifying-and-improving-on-policy-distillation-for-few-step-diffusion-2610-18472?lang=esChiosaes2026-10-14T11:47:52ZUn método de regresión destila modelos generativos en menos pasoshttps://chiosa.app/papers/active-noise-cancellation-extends-the-benefit-of-ai-speech-enhancement-in-real-hearing-aid-recordings-2610-17287Chiosaen2026-10-14T11:47:48ZANC lets hearing-aid speech enhancement reach lower frequencieshttps://chiosa.app/papers/active-noise-cancellation-extends-the-benefit-of-ai-speech-enhancement-in-real-hearing-aid-recordings-2610-17287?lang=itChiosait2026-10-14T11:47:48ZCon ANC, il miglioramento del parlato raggiunge frequenze più bassehttps://chiosa.app/papers/active-noise-cancellation-extends-the-benefit-of-ai-speech-enhancement-in-real-hearing-aid-recordings-2610-17287?lang=esChiosaes2026-10-14T11:47:48ZLa cancelación de ruido activa permite que la mejora de voz llegue a frecuencias más bajas en audífonoshttps://chiosa.app/papers/active-noise-cancellation-extends-the-benefit-of-ai-speech-enhancement-in-real-hearing-aid-recordings-2610-17287?lang=frChiosafr2026-10-14T11:47:48ZL’ANC étend l’amélioration de la parole des aides auditives vers les basses fréquenceshttps://chiosa.app/papers/aegisasync-symbolic-async-event-simulation-for-risks-in-python-asynchronous-programs-2610-18831Chiosaen2026-10-14T11:47:48ZAegisAsync traces Python async behavior to flag riskshttps://chiosa.app/papers/aegisasync-symbolic-async-event-simulation-for-risks-in-python-asynchronous-programs-2610-18831?lang=itChiosait2026-10-14T11:47:48ZAegisAsync traccia il comportamento async in Python per rilevare i rischihttps://chiosa.app/papers/aegisasync-symbolic-async-event-simulation-for-risks-in-python-asynchronous-programs-2610-18831?lang=esChiosaes2026-10-14T11:47:48ZAegisAsync rastrea el comportamiento asíncrono de Python para detectar riesgoshttps://chiosa.app/papers/aegisasync-symbolic-async-event-simulation-for-risks-in-python-asynchronous-programs-2610-18831?lang=frChiosafr2026-10-14T11:47:48ZAegisAsync suit le comportement asynchrone en Python pour repérer des risqueshttps://chiosa.app/papers/attune-agentic-throughput-tuning-for-llm-training-2610-17528Chiosaen2026-10-14T11:47:48ZA coding agent can tune training recipes while checks guard correctnesshttps://chiosa.app/papers/attune-agentic-throughput-tuning-for-llm-training-2610-17528?lang=itChiosait2026-10-14T11:47:48ZUn agente di coding ottimizza le configurazioni di addestramento con controlli di correttezzahttps://chiosa.app/papers/attune-agentic-throughput-tuning-for-llm-training-2610-17528?lang=esChiosaes2026-10-14T11:47:48ZUn agente de programación puede ajustar recetas de entrenamiento con controles de validezhttps://chiosa.app/papers/attune-agentic-throughput-tuning-for-llm-training-2610-17528?lang=frChiosafr2026-10-14T11:47:48ZUn agent de programmation peut régler les recettes d’entraînement sous contrôle de validitéhttps://chiosa.app/papers/bifocal-training-free-mixed-granularity-visual-token-pruning-for-multimodal-llms-2610-18399Chiosaen2026-10-14T11:47:48ZBifocal preserves image detail and context while trimming visual inputhttps://chiosa.app/papers/bifocal-training-free-mixed-granularity-visual-token-pruning-for-multimodal-llms-2610-18399?lang=itChiosait2026-10-14T11:47:48ZBifocal conserva dettagli e contesto riducendo l’input visivohttps://chiosa.app/papers/bifocal-training-free-mixed-granularity-visual-token-pruning-for-multimodal-llms-2610-18399?lang=esChiosaes2026-10-14T11:47:48ZBifocal conserva detalle y contexto al reducir la entrada visualhttps://chiosa.app/papers/bifocal-training-free-mixed-granularity-visual-token-pruning-for-multimodal-llms-2610-18399?lang=frChiosafr2026-10-14T11:47:48ZBifocal préserve les détails et le contexte de l’image en réduisant les unités visuelleshttps://chiosa.app/papers/calqa-conformal-screening-of-forecast-ensemble-configurations-from-one-mergeable-state-2610-13249Chiosaen2026-10-14T11:47:48ZA compact screen speeds forecast-ensemble searcheshttps://chiosa.app/papers/calqa-conformal-screening-of-forecast-ensemble-configurations-from-one-mergeable-state-2610-13249?lang=itChiosait2026-10-14T11:47:48ZUn riepilogo compatto accelera la ricerca sugli ensemble di previsionihttps://chiosa.app/papers/calqa-conformal-screening-of-forecast-ensemble-configurations-from-one-mergeable-state-2610-13249?lang=esChiosaes2026-10-14T11:47:48ZUn resumen compacto acelera la búsqueda de pronósticos combinadoshttps://chiosa.app/papers/calqa-conformal-screening-of-forecast-ensemble-configurations-from-one-mergeable-state-2610-13249?lang=frChiosafr2026-10-14T11:47:48ZUn filtre compact accélère la recherche de poids pour les prévisionshttps://chiosa.app/papers/compositional-rl-evaluating-and-training-llms-for-collaborative-reasoning-2610-17456Chiosaen2026-10-14T11:47:48ZSolo scores often miss which models help a teamhttps://chiosa.app/papers/compositional-rl-evaluating-and-training-llms-for-collaborative-reasoning-2610-17456?lang=itChiosait2026-10-14T11:47:48ZI punteggi individuali spesso non rivelano quali modelli aiutano il gruppohttps://chiosa.app/papers/compositional-rl-evaluating-and-training-llms-for-collaborative-reasoning-2610-17456?lang=esChiosaes2026-10-14T11:47:48ZLa puntuación individual suele ocultar qué modelos ayudan al equipohttps://chiosa.app/papers/compositional-rl-evaluating-and-training-llms-for-collaborative-reasoning-2610-17456?lang=frChiosafr2026-10-14T11:47:48ZLes scores individuels indiquent mal quels modèles aident l’équipehttps://chiosa.app/papers/dpoison-vl-clean-preference-poisoning-in-vision-language-direct-preference-optimization-2610-18609Chiosaen2026-10-14T11:47:48ZSmall image changes can steer preference traininghttps://chiosa.app/papers/dpoison-vl-clean-preference-poisoning-in-vision-language-direct-preference-optimization-2610-18609?lang=esChiosaes2026-10-14T11:47:48ZPequeños cambios en imágenes pueden orientar el entrenamiento por preferenciashttps://chiosa.app/papers/dpoison-vl-clean-preference-poisoning-in-vision-language-direct-preference-optimization-2610-18609?lang=frChiosafr2026-10-14T11:47:48ZDe légères modifications d’image peuvent orienter l’entraînement par préférenceshttps://chiosa.app/papers/enabling-token-wise-importance-sampling-in-off-policy-rlvr-with-proximal-kl-regularization-2610-18738Chiosaen2026-10-14T11:47:48ZA token-wise update keeps stale-response training aimed at its targethttps://chiosa.app/papers/enabling-token-wise-importance-sampling-in-off-policy-rlvr-with-proximal-kl-regularization-2610-18738?lang=itChiosait2026-10-14T11:47:48ZUn aggiornamento per token mantiene l'obiettivo con risposte precedentihttps://chiosa.app/papers/enabling-token-wise-importance-sampling-in-off-policy-rlvr-with-proximal-kl-regularization-2610-18738?lang=esChiosaes2026-10-14T11:47:48ZUna actualización por token mantiene el objetivo al entrenar con respuestas antiguashttps://chiosa.app/papers/enabling-token-wise-importance-sampling-in-off-policy-rlvr-with-proximal-kl-regularization-2610-18738?lang=frChiosafr2026-10-14T11:47:48ZUne mise à jour token par token maintient l'entraînement sur réponses obsolètes orienté vers sa ciblehttps://chiosa.app/papers/evidence-first-agentic-verification-2610-17724Chiosaen2026-10-14T11:47:48ZSauron checks evidence before judging agent taskshttps://chiosa.app/papers/evidence-first-agentic-verification-2610-17724?lang=itChiosait2026-10-14T11:47:48ZSauron verifica le prove prima di giudicare le attività degli agentihttps://chiosa.app/papers/evidence-first-agentic-verification-2610-17724?lang=esChiosaes2026-10-14T11:47:48ZSauron verifica pruebas antes de evaluar tareas de agenteshttps://chiosa.app/papers/evidence-first-agentic-verification-2610-17724?lang=frChiosafr2026-10-14T11:47:48ZSauron vérifie les preuves avant de juger les tâches agentshttps://chiosa.app/papers/fan-out-or-relay-ownership-aware-routing-for-llm-agents-querying-siloed-tables-2610-18157Chiosaen2026-10-14T11:47:48ZA schema draft helps route questions across private data ownershttps://chiosa.app/papers/fan-out-or-relay-ownership-aware-routing-for-llm-agents-querying-siloed-tables-2610-18157?lang=itChiosait2026-10-14T11:47:48ZUna bozza basata sugli schemi guida le domande tra proprietari di dati privatihttps://chiosa.app/papers/fan-out-or-relay-ownership-aware-routing-for-llm-agents-querying-siloed-tables-2610-18157?lang=esChiosaes2026-10-14T11:47:48ZUn borrador de consulta ayuda a dirigir preguntas entre propietarios de datos privadoshttps://chiosa.app/papers/fan-out-or-relay-ownership-aware-routing-for-llm-agents-querying-siloed-tables-2610-18157?lang=frChiosafr2026-10-14T11:47:48ZUn brouillon de schéma aide à orienter les questions vers des détenteurs de données privées