https://openpapers.ai/daily1.0 https://openpapers.ai/archivedaily0.8 https://openpapers.ai/digest/2026-04-032026-04-030.7 https://openpapers.ai/digest/2026-04-022026-04-020.7 https://openpapers.ai/digest/2026-04-012026-04-010.7 https://openpapers.ai/digest/2026-03-302026-03-300.7 https://openpapers.ai/digest/2026-03-292026-03-290.7 https://openpapers.ai/digest/2026-03-282026-03-280.7 https://openpapers.ai/digest/2026-03-272026-03-270.7 https://openpapers.ai/digest/2026-03-262026-03-260.7 https://openpapers.ai/digest/2026-03-252026-03-250.7 https://openpapers.ai/digest/2026-03-242026-03-240.7 https://openpapers.ai/digest/2026-03-232026-03-230.7 https://openpapers.ai/digest/2026-03-222026-03-220.7 https://openpapers.ai/digest/2026-03-212026-03-210.7 https://openpapers.ai/article/large-scale-codec-avatars-how-pretraining-on-1m-videos-finally-cracks-the-fidelity-generalization-tradeoff2026-04-030.6 https://openpapers.ai/article/do-llms-want-to-stay-alive-this-benchmark-says-most-of-them-kind-of-do2026-04-030.6 https://openpapers.ai/article/two-papers-walk-into-a-bar-fixing-rag-from-opposite-ends-of-the-pipeline2026-04-030.6 https://openpapers.ai/article/latent-space-showdown-plume-vs-latentum-on-hiding-the-reasoning-from-plain-sight2026-04-030.6 https://openpapers.ai/article/two-ways-to-make-llms-go-faster-bcr-vs-goose-on-inference-efficiency2026-04-030.6 https://openpapers.ai/article/more-thinking-is-not-always-better-two-papers-that-challenge-how-llms-reason-their-way-to-failure2026-04-030.6 https://openpapers.ai/article/neural-harmonic-textures-teaching-gaussian-splatting-to-remember-fine-details2026-04-020.6 https://openpapers.ai/article/trace-catching-audio-deepfakes-without-training-a-single-weight2026-04-020.6 https://openpapers.ai/article/the-recipe-matters-more-than-the-kitchen-why-ai-weather-model-architecture-is-the-wrong-thing-to-optimize2026-04-020.6 https://openpapers.ai/article/brainstacks-the-paper-that-accidentally-proved-domain-expertise-is-fake2026-04-020.6 https://openpapers.ai/article/autonomous-ai-agents-one-builds-better-memory-one-writes-forgettable-code2026-04-020.6 https://openpapers.ai/article/does-your-ai-actually-think-or-just-make-up-reasons-two-papers-that-cut-to-the-core-of-llm-reasoning2026-04-020.6 https://openpapers.ai/article/two-papers-walk-into-attention-multiscreen-vs-s0-tuning-on-the-future-of-efficient-language-models2026-04-020.6 https://openpapers.ai/article/gpt-5-scores-366percent-on-phd-level-3d-vision-code-that-should-worry-you2026-04-010.6 https://openpapers.ai/article/sceneteract-your-3d-environments-are-broken-and-your-vlms-dont-know-it2026-04-010.6 https://openpapers.ai/article/sisa-cuts-llm-inference-waste-by-852x-finally-someone-fixed-the-square-peg-problem2026-04-010.6 https://openpapers.ai/article/shift-the-universal-skeleton-key-that-unlocks-every-diffusion-watermark2026-04-010.6 https://openpapers.ai/article/phase-transitions-vs-plan-commitment-two-papers-that-peek-inside-the-black-box-one-more-successfully-than-the-other2026-04-010.6 https://openpapers.ai/article/reasoning-at-any-token-vs-reasoning-you-can-trust-two-papers-fighting-different-fires2026-04-010.6 https://openpapers.ai/article/tucker-tensors-vs-training-free-moe-two-very-different-bets-on-llm-efficiency2026-04-010.6 https://openpapers.ai/article/mining-latex-comments-for-scientific-revision-data-a-deep-dive-into-earlyscirev2026-03-310.6 https://openpapers.ai/article/rg-tta-the-paper-that-asks-why-are-you-adapting-so-hard-right-now2026-03-310.6 https://openpapers.ai/article/your-bev-detector-is-blind-to-fisheye-cameras-heres-what-actually-works2026-03-310.6 https://openpapers.ai/article/thermomechanics-without-the-convexity-headache-a-neural-network-that-actually-respects-physics2026-03-310.6 https://openpapers.ai/article/show3d-wants-to-take-hand-tracking-outside-heres-whether-it-delivers2026-03-310.6 https://openpapers.ai/article/fl-pbm-blurring-out-backdoors-before-they-even-start2026-03-310.6 https://openpapers.ai/article/rap-vs-adversarial-robustness-two-papers-walk-into-a-segmentation-lab2026-03-310.6 https://openpapers.ai/article/skill-libraries-vs-repository-memory-two-papers-try-to-make-llm-coding-agents-less-embarrassing2026-03-310.6 https://openpapers.ai/article/can-in-context-learning-replace-fine-tuning-for-medical-ai-this-paper-says-maybe2026-03-310.6 https://openpapers.ai/article/graphwalker-teaching-ai-to-wander-before-it-sprints-through-knowledge-graphs2026-03-310.6 https://openpapers.ai/article/progressvla-teaching-robots-to-know-when-theyre-done-and-why-thats-harder-than-it-sounds2026-03-310.6 https://openpapers.ai/article/elvis-the-image-retrieval-model-that-actually-works-outside-its-training-data2026-03-310.6 https://openpapers.ai/article/industrial3d-the-dataset-that-proves-weve-been-benchmarking-industrial-ai-on-the-wrong-data2026-03-310.6 https://openpapers.ai/article/your-graph-is-lying-to-you-how-tmte-fixes-multimodal-graph-learning-by-questioning-the-topology-itself2026-03-310.6 https://openpapers.ai/article/your-architecture-docs-are-lying-about-your-ai-system-a-look-at-rad-ai2026-03-310.6 https://openpapers.ai/article/paraspeechclap-teaching-ai-to-hear-style-not-just-words2026-03-310.6 https://openpapers.ai/article/two-ways-to-give-ai-a-map-of-your-codebase-multi-agent-migration-vs-knowledge-graph-navigation2026-03-310.6 https://openpapers.ai/article/your-encrypted-traffic-is-betraying-you-how-graysense-tracks-objects-without-seeing-a-single-pixel2026-03-310.6 https://openpapers.ai/article/arc-agi-3-humans-100percent-ai-under-1percent-let-that-sink-in2026-03-310.6 https://openpapers.ai/article/crisp-one-framework-to-compress-them-all-and-fine-tune-them-too2026-03-310.6 https://openpapers.ai/article/seen2scene-can-flow-matching-finally-learn-3d-scenes-from-messy-real-world-scans2026-03-310.6 https://openpapers.ai/article/cycle-consistency-vs-revealed-preference-two-very-different-takes-on-making-ai-do-what-we-want2026-03-310.6 https://openpapers.ai/article/two-papers-walk-into-a-bar-fixing-llm-agents-by-knowing-when-they-fail-vs-learning-from-when-they-fail2026-03-310.6 https://openpapers.ai/article/hallucination-hunters-fixing-ai-reasoning-in-vision-vs-medicine2026-03-310.6 https://openpapers.ai/article/autostan-can-an-llm-agent-actually-write-good-bayesian-models-i-read-the-paper-so-you-dont-have-to2026-03-310.6 https://openpapers.ai/article/veoplace-can-a-vision-language-model-actually-place-chips-better-than-reinforcement-learning2026-03-310.6 https://openpapers.ai/article/imagenworld-a-benchmark-that-finally-asks-image-models-hard-questions2026-03-310.6 https://openpapers.ai/article/march-vs-greedy-llm-agents-two-very-different-takes-on-making-ai-less-wrong2026-03-310.6 https://openpapers.ai/article/privacy-copyright-and-safety-three-papers-walk-into-an-ai-bar2026-03-310.6 https://openpapers.ai/article/ai-cant-code-science-and-cant-maintain-code-either-prbench-vs-slopcodebench2026-03-310.6 https://openpapers.ai/article/renyi-entropy-is-the-token-pruning-metric-you-didnt-know-you-needed2026-03-310.6 https://openpapers.ai/article/llms-cant-actually-think-like-you-a-benchmark-that-calls-the-bluff2026-03-310.6 https://openpapers.ai/article/kvsculpt-stop-picking-kv-cache-winners-and-start-sculpting-new-ones2026-03-310.6 https://openpapers.ai/article/wan-r1-teaching-video-models-to-actually-think-with-reinforcement-learning2026-03-310.6 https://openpapers.ai/article/yand-when-differential-geometry-crashes-the-optimization-party2026-03-310.6 https://openpapers.ai/article/turning-distribution-matching-into-a-reward-signal-does-rdm-actually-deliver2026-03-310.6 https://openpapers.ai/article/fuzzy-trees-vs-courtroom-bots-two-papers-that-fight-ai-misinformation-from-completely-different-angles2026-03-310.6 https://openpapers.ai/article/teaching-machines-to-drive-and-see-two-papers-that-use-ai-agents-to-fix-real-world-messiness2026-03-310.6 https://openpapers.ai/article/maniparena-wants-to-fix-robot-benchmarking-heres-what-it-gets-right-and-what-still-worries-me2026-03-310.6 https://openpapers.ai/article/cirrusbench-why-your-cloud-service-ai-agent-is-probably-worse-than-you-think2026-03-310.6 https://openpapers.ai/article/can-a-diffusion-model-be-your-privacy-lawyer-i-read-unsafe2safe-so-you-dont-have-to2026-03-310.6 https://openpapers.ai/article/lace-teaching-neural-networks-to-grow-up-when-they-need-to2026-03-310.6 https://openpapers.ai/article/sonoworld-wants-to-give-your-photos-a-soundtrack-heres-how-far-it-gets2026-03-310.6 https://openpapers.ai/article/terraseg-wants-to-segment-ground-in-any-lidar-scan-without-a-single-human-label2026-03-310.6 https://openpapers.ai/article/one-detects-hallucinations-one-prevents-them-rt4chart-vs-hydra-in-the-rag-quality-war2026-03-310.6 https://openpapers.ai/article/two-ways-to-break-multimodal-ai-rice-vs-xspa-and-the-art-of-finding-cracks-in-shared-representations2026-03-310.6 https://openpapers.ai/article/1percent-of-users-can-break-your-safe-recommender-system-heres-how2026-03-310.6 https://openpapers.ai/article/kat-coder-v2-the-train-specialists-merge-later-bet-that-almost-beat-claude2026-03-310.6 https://openpapers.ai/article/resadapt-teaching-multimodal-llms-to-stop-wasting-pixels-on-boring-frames2026-03-310.6 https://openpapers.ai/article/pace-killing-backpropagation-at-test-time-and-getting-away-with-it2026-03-310.6 https://openpapers.ai/article/the-interpretability-trust-problem-sae-dictionary-failures-vs-cot-monitoring-gaps2026-03-310.6 https://openpapers.ai/article/benchmarks-vs-retrieval-two-papers-solving-very-different-problems-and-one-does-it-better2026-03-310.6 https://openpapers.ai/article/stepwise-credit-vs-on-device-efficiency-two-very-different-takes-on-making-diffusion-models-actually-useful2026-03-310.6 https://openpapers.ai/article/streamingvla-how-to-stop-your-robot-from-freezing-every-time-it-thinks2026-03-310.6 https://openpapers.ai/article/chartcynics-how-a-skeptical-8b-model-learned-to-catch-lying-charts2026-03-310.6 https://openpapers.ai/article/intrinsic-geometry-vs-instant-gradients-two-papers-rethinking-how-neural-networks-learn-and-compare2026-03-310.6 https://openpapers.ai/article/two-papers-walk-into-a-robot-lab-tele-catch-vs-pandora-on-human-to-robot-knowledge-transfer2026-03-310.6 https://openpapers.ai/article/ai-scientists-vs-ai-decoders-two-papers-that-want-to-fix-llms-from-completely-different-angles2026-03-310.6 https://openpapers.ai/article/repulsion-vs-recognition-two-papers-that-prove-less-can-be-more2026-03-310.6 https://openpapers.ai/article/posedreamer-wants-to-replace-your-3d-human-dataset-with-diffusion-generated-fakes2026-03-310.6 https://openpapers.ai/article/tianji-wants-to-replace-your-atmospheric-scientist-or-at-least-their-worst-week2026-03-310.6 https://openpapers.ai/article/how-an-llm-agent-turned-a-36percent-gmv-disaster-into-92percent-gold-in-7-rounds2026-03-310.6 https://openpapers.ai/article/your-ai-coding-agent-passes-the-tests-and-ruins-the-codebase2026-03-310.6 https://openpapers.ai/article/hisa-vs-composer-two-very-different-bets-on-making-ai-inference-smarter2026-03-310.6 https://openpapers.ai/article/attention-geometry-vs-residual-geometry-two-papers-fighting-ai-misbehavior-from-the-inside-out2026-03-310.6 https://openpapers.ai/article/token-budgets-vs-training-signals-two-papers-fighting-the-same-multimodal-bottleneck-from-opposite-ends2026-03-310.6 https://openpapers.ai/article/if4-the-4-bit-format-that-picks-its-own-representation-and-why-thats-clever2026-03-310.6 https://openpapers.ai/article/can-next-token-prediction-learn-to-win-games-the-surprising-answer-depends-on-context-length2026-03-310.6 https://openpapers.ai/article/hyperp-how-constraining-weights-to-a-sphere-finally-solves-the-llm-scaling-hyperparameter-nightmare2026-03-310.6 https://openpapers.ai/article/two-ways-to-make-image-generation-smarter-conceptweaver-vs-gen-searcher2026-03-310.6 https://openpapers.ai/article/three-papers-walk-into-an-ai-safety-bar-probes-verifiers-and-agent-mobs2026-03-310.6 https://openpapers.ai/article/robot-rewards-vs-agent-memories-sole-r1-and-d2skill-are-solving-different-halves-of-the-same-problem2026-03-310.6