https://chiosa.app/papers/autocompress-automated-hardware-guided-model-compression-2611-08290Chiosaen2026-11-06T20:02:35ZHardware-aware pruning finds models that fit serving limitshttps://chiosa.app/papers/evomemstate-evolving-memory-state-for-personal-agents-2611-03743Chiosaen2026-11-06T20:02:35ZPersonal agents learn how to use evolving memorieshttps://chiosa.app/papers/interactive-visual-gym-evaluating-visual-interaction-models-with-simulated-users-in-interactive-worlds-2611-08286Chiosaen2026-11-06T20:02:35ZSimulated user reactions expose limits in visual assistancehttps://chiosa.app/papers/on-the-opportunities-and-challenges-of-prefill-to-autoregressive-cross-model-transfer-2611-06976Chiosaen2026-11-06T20:02:35ZPLatour links larger prompt models to smaller generatorshttps://chiosa.app/papers/sla-scaling-spatial-language-action-model-via-continual-pre-training-2611-08373Chiosaen2026-11-06T20:02:35ZRobot training aligns 2D vision with 3D spatial knowledgehttps://chiosa.app/papers/the-intrusion-pathway-how-non-evidential-context-degrades-language-model-performance-2611-07778Chiosaen2026-11-06T20:02:35ZIrrelevant text weakens answers while subject information remainshttps://chiosa.app/papers/a-behavioural-framework-for-predicting-and-understanding-loss-of-control-in-frontier-artificial-intelligence-2611-07953Chiosaen2026-11-06T15:39:43ZA three-factor framework links warning signs to simulated control failureshttps://chiosa.app/papers/a-gingerbread-house-for-llm-agents-source-planned-honeypots-that-lure-autonomous-attackers-into-an-isolated-s-2611-07611Chiosaen2026-11-06T15:39:43ZA honeypot routes confirmed web attackers to a copyhttps://chiosa.app/papers/aixapply-fast-and-reliable-full-file-code-integration-with-specialized-small-models-for-ide-workflows-2611-07634Chiosaen2026-11-06T15:39:43ZA specialized 4B model can integrate code snippets accuratelyhttps://chiosa.app/papers/benchmarks-as-software-terminal-bench-and-the-case-for-continuous-benchmarks-2611-08417Chiosaen2026-11-06T15:39:43ZA continuing audit finds defects in released agent benchmarkshttps://chiosa.app/papers/emoreward-self-evolving-reward-calibration-for-emotional-video-generation-rl-2611-06518Chiosaen2026-11-06T15:39:43ZA reward system learns which emotional video cues to trusthttps://chiosa.app/papers/fastafd-when-does-attention-ffn-disaggregation-pay-on-rack-scale-nvlink-2611-07149Chiosaen2026-11-06T15:39:43ZSeparating expert work helps some rack-scale language modelshttps://chiosa.app/papers/freeocc-decoupling-ego-motion-for-efficient-4d-occupancy-forecasting-via-continuous-flow-matching-2611-08128Chiosaen2026-11-06T15:39:43ZFreeOcc uses a planned route to forecast driving sceneshttps://chiosa.app/papers/from-pixels-to-actions-towards-native-vision-language-action-policies-2611-06918Chiosaen2026-11-06T15:39:43ZRobot policies improve when they can read earlier visual featureshttps://chiosa.app/papers/frontierphysics-evaluating-agents-for-end-to-end-physics-research-2611-06910Chiosaen2026-11-06T15:39:43ZA new physics test finds research remains hard for AI agentshttps://chiosa.app/papers/gabor-gabor-e-gaborformer-a-strictly-causal-speech-recognizer-with-gabor-dictionary-convolutions-and-delta-ru-2611-08092Chiosaen2026-11-06T15:39:43ZA speech encoder keeps streaming memory boundedhttps://chiosa.app/papers/gameverse-a-minute-scale-gameplay-dataset-for-long-horizon-interactive-world-modeling-2610-37257Chiosaen2026-11-06T15:39:43ZGameVerse supplies two-minute gameplay clips for world modelshttps://chiosa.app/papers/harness-s-learning-from-sibling-harnesses-for-generalizable-agent-training-2611-04277Chiosaen2026-11-06T15:39:43ZHarness-S transfers sibling skills into one agenthttps://chiosa.app/papers/kinematic-hypothesis-context-embodiment-conditioned-learned-video-coding-for-bandwidth-limited-robot-policy-u-2611-07411Chiosaen2026-11-06T15:39:43ZRobot motion helps video codecs save bitshttps://chiosa.app/papers/on-the-generalization-limitation-of-on-policy-distillation-for-post-rl-consolidation-2611-05609Chiosaen2026-11-06T15:39:43ZDistillation can miss abilities beyond its training exampleshttps://chiosa.app/papers/per-query-recall-guarantees-for-approximate-nearest-neighbor-search-in-dynamic-vector-databases-2611-07579Chiosaen2026-11-06T15:39:43ZPQR-Search bounds recall for individual querieshttps://chiosa.app/papers/pyrua-datafarm-llm-agent-as-a-cost-efficient-teleoperator-for-real-ready-robot-data-2611-05320Chiosaen2026-11-06T15:39:43ZA robot data system makes simulator motion look like real sceneshttps://chiosa.app/papers/random-is-almost-enough-stable-and-low-cost-benchmark-compression-via-bestrand-2611-00883Chiosaen2026-11-06T15:39:43ZBestRand selects smaller tests to steady model scoreshttps://chiosa.app/papers/recursive-sparse-attention-million-token-length-extrapolation-at-logarithmic-cost-2611-07232Chiosaen2026-11-06T15:39:43ZA tree-based method extends language models to million token contextshttps://chiosa.app/papers/signed-but-not-safe-requirement-integrity-in-multi-agent-llm-handoffs-2611-07483Chiosaen2026-11-06T15:39:43ZSigned handoffs reduce violations in agent chainshttps://chiosa.app/papers/streamduo-a-live-commentator-and-a-research-agent-for-unseen-games-2611-01120Chiosaen2026-11-06T15:39:43ZA game commentator learns when to ask for unfamiliar nameshttps://chiosa.app/papers/tok-a-textual-concrete-syntax-of-bpmn-and-dmn-for-humans-and-ai-agents-2611-07945Chiosaen2026-11-06T15:39:43ZA checked text format for process modelshttps://chiosa.app/papers/triattention-unified-sparse-mixed-precision-attention-for-long-context-llm-inference-2611-07256Chiosaen2026-11-06T15:39:43ZAttention gets cheaper by matching precision to each regionhttps://chiosa.app/papers/turns-matter-advancing-streaming-video-rl-via-turn-level-credit-2611-06728Chiosaen2026-11-06T15:39:43ZTurnWise gives streaming video decisions credit turn by turnhttps://chiosa.app/papers/unpacking-the-compute-efficiency-of-strong-to-weak-opd-against-grpo-2611-08453Chiosaen2026-11-06T15:39:43ZDistillation saves compute when it merges trained specialistshttps://chiosa.app/papers/vissearch-bench-evaluating-long-horizon-visual-search-in-multimodal-agents-2611-04894Chiosaen2026-11-06T15:39:43ZVisSearch-Bench tests how agents connect images and searcheshttps://chiosa.app/papers/welm-image-a-unified-image-generation-model-towards-visual-intelligence-2611-07392Chiosaen2026-11-06T15:39:43ZOne image model handles creation, editing and referenceshttps://chiosa.app/papers/what-does-multi-turn-reinforcement-learning-really-improve-in-llm-agents-2611-06483Chiosaen2026-11-06T15:39:43ZMulti-turn RL gains depend on interaction budgetshttps://chiosa.app/papers/where-do-reported-long-video-qa-gains-come-from-a-controlled-audit-of-frame-selection-evidence-injection-and-2611-07329Chiosaen2026-11-06T15:39:43ZFrame choice had bounded gains in a long-video audithttps://chiosa.app/papers/does-your-generalist-skill-optimizer-really-incentivize-agentic-tasks-rethinking-skill-optimization-via-bipar-2610-37201Chiosaen2026-11-06T15:34:40ZLearned agent skills added little test success and used more tokens