Every paper checked — nothing hidden

All papers

Every paper checked for brief #4 of 12–19 Oct 2026 — nothing is hidden. Brief papers first.

All17,313 In brief24 More (on-topic)3,333 Not relevant13,754 Not assessed87 Older352 Could not confirm93 Blogs with no paper266 All blogs284

Paper	Topic	Relevance
MIDI-REPA: Improving Text-to-Music Generation with Symbolic Representation Alignment The paper aligns intermediate representations of text-to-music models with symbolic music transcriptions to improve generation quality efficiently. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Music generation alignment technique, not model compression or serving efficiency.
MIDI-REPA: Improving Text-to-Music Generation with Symbolic Representation Alignment The paper aligns intermediate representations of text-to-music models with symbolic music transcriptions to improve generation quality efficiently. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Addresses audio generation quality, not safety, guardrails, or hallucination.
Two-Stage Score Estimation for Group-Specific Diffusion Models The paper establishes theoretical convergence rates for group-specific diffusion models that learn a shared score alongside a target-specific offset. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Theoretical score estimation framework, not efficient model serving or compression.
Two-Stage Score Estimation for Group-Specific Diffusion Models The paper establishes theoretical convergence rates for group-specific diffusion models that learn a shared score alongside a target-specific offset. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Studies theoretical density estimation, not language model benchmark evaluation.
Two-Stage Score Estimation for Group-Specific Diffusion Models The paper establishes theoretical convergence rates for group-specific diffusion models that learn a shared score alongside a target-specific offset. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Focuses on mathematical diffusion bounds, not AI safety or jailbreaks.
The Blume-Capel Model as a $φ^6$ Field Theory: A Monte Carlo and Flow Model Study The paper evaluates continuous spin Blume-Capel models using Monte Carlo simulations and normalizing flows to accurately map phase diagrams. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Statistical physics sampling paper, not LLM evaluation or benchmarks.
The Blume-Capel Model as a $φ^6$ Field Theory: A Monte Carlo and Flow Model Study The paper evaluates continuous spin Blume-Capel models using Monte Carlo simulations and normalizing flows to accurately map phase diagrams. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Studies physical spin systems, not AI safety or deployment.
The Blume-Capel Model as a $φ^6$ Field Theory: A Monte Carlo and Flow Model Study The paper evaluates continuous spin Blume-Capel models using Monte Carlo simulations and normalizing flows to accurately map phase diagrams. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Focuses on physics simulation methods, not efficient model inference.
Halve-Then-Greedy: Adaptive Sampling for Posterior Success with Near-Optimal Instance Dependence The paper presents an adaptive bandit sampling algorithm designed to verify that an option exceeds a performance threshold within a fixed budget. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on multi-armed bandit sampling theory, not LLM benchmarks.
Halve-Then-Greedy: Adaptive Sampling for Posterior Success with Near-Optimal Instance Dependence The paper presents an adaptive bandit sampling algorithm designed to verify that an option exceeds a performance threshold within a fixed budget. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Theoretical bandit sampling analysis, not deployment safety or guardrails.
Halve-Then-Greedy: Adaptive Sampling for Posterior Success with Near-Optimal Instance Dependence The paper presents an adaptive bandit sampling algorithm designed to verify that an option exceeds a performance threshold within a fixed budget. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Bandit sampling optimization theory, not small or efficient models.
Imperfect Is Perfect: Learned Column Prefilling for Branch-Price-and-Cut The paper uses neural networks to predict initial columns and dual centers, speeding up branch-price-and-cut solvers for vehicle routing problems. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Studies discrete optimization algorithms, not AI safety or reliability.
Imperfect Is Perfect: Learned Column Prefilling for Branch-Price-and-Cut The paper uses neural networks to predict initial columns and dual centers, speeding up branch-price-and-cut solvers for vehicle routing problems. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Focuses on vehicle routing optimization, not efficient neural models.
Imperfect Is Perfect: Learned Column Prefilling for Branch-Price-and-Cut The paper uses neural networks to predict initial columns and dual centers, speeding up branch-price-and-cut solvers for vehicle routing problems. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Addresses operations research optimization, not LLM evaluation methods.
CausalEvo: Causal Skill Self-Evolution for Health LLM Agents The paper introduces a causal evaluation method for healthcare language model agents to iteratively select and retain useful task skills. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Focuses on agent skill optimization, not model compression or serving.
CausalEvo: Causal Skill Self-Evolution for Health LLM Agents The paper introduces a causal evaluation method for healthcare language model agents to iteratively select and retain useful task skills. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on agent skill evolution, not benchmark methodology or contamination.
Acceleration and its Limits in Gaussian Variational Inference The paper studies query complexity bounds and limits of momentum-based acceleration in Gaussian variational inference for strongly log-concave distributions. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on variational inference theory, not LLM benchmark evaluation.
Acceleration and its Limits in Gaussian Variational Inference The paper studies query complexity bounds and limits of momentum-based acceleration in Gaussian variational inference for strongly log-concave distributions. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Focuses on optimization complexity, not AI safety or deployment robustness.
Acceleration and its Limits in Gaussian Variational Inference The paper studies query complexity bounds and limits of momentum-based acceleration in Gaussian variational inference for strongly log-concave distributions. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Addresses optimization theory in statistics, not efficient model architectures.
On the Last-Iterate Convergence of Clipped Gradient Methods The paper provides theoretical last-iterate convergence guarantees for clipped stochastic gradient methods under heavy-tailed noise. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Theoretical analysis of optimization bounds, not model efficiency or serving.
On the Last-Iterate Convergence of Clipped Gradient Methods The paper provides theoretical last-iterate convergence guarantees for clipped stochastic gradient methods under heavy-tailed noise. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on theoretical optimization convergence, not LLM evaluation benchmarks.
On the Last-Iterate Convergence of Clipped Gradient Methods The paper provides theoretical last-iterate convergence guarantees for clipped stochastic gradient methods under heavy-tailed noise. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Studies gradient descent math, not safety, guardrails, or alignment.
SALTO: Solver-Anchored Lifted Transfer Operators for Nonlinear Event Inference The paper introduces an operator-learning framework using linear transfer operators to infer localized finite-amplitude source events in nonlinear dynamical systems. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Scientific inverse problem modeling, not AI safety, jailbreak defense, or guardrails.
SALTO: Solver-Anchored Lifted Transfer Operators for Nonlinear Event Inference The paper introduces an operator-learning framework using linear transfer operators to infer localized finite-amplitude source events in nonlinear dynamical systems. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Dynamical transfer operator method, not model quantization, distillation, or efficient serving.
SALTO: Solver-Anchored Lifted Transfer Operators for Nonlinear Event Inference The paper introduces an operator-learning framework using linear transfer operators to infer localized finite-amplitude source events in nonlinear dynamical systems. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Nonlinear dynamical system inference paper, not LLM benchmark evaluation.
Score Mismatch and Set-Constrained Sampling in Zero-Shot Conditional Discrete Diffusion The paper provides theoretical bounds on the approximation error caused by score mismatch during zero-shot conditional sampling in discrete diffusion models. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Focuses on discrete diffusion distribution shifts, not deployment safety or guardrails.
Score Mismatch and Set-Constrained Sampling in Zero-Shot Conditional Discrete Diffusion The paper provides theoretical bounds on the approximation error caused by score mismatch during zero-shot conditional sampling in discrete diffusion models. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Theoretical discrete diffusion analysis, not model compression, distillation, or serving efficiency.
Score Mismatch and Set-Constrained Sampling in Zero-Shot Conditional Discrete Diffusion The paper provides theoretical bounds on the approximation error caused by score mismatch during zero-shot conditional sampling in discrete diffusion models. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Theoretical analysis of diffusion sampling bias, not LLM benchmark evaluation.
Adaptive Randomized Neural Networks for PDEs: Structure-Aware Feature Redistribution and Residual-Driven Block Enrichment The paper introduces a two-stage adaptive feature construction method to improve randomized neural networks for solving partial differential equations. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Randomized neural network for PDEs, not efficient language model serving or compression.
Adaptive Randomized Neural Networks for PDEs: Structure-Aware Feature Redistribution and Residual-Driven Block Enrichment The paper introduces a two-stage adaptive feature construction method to improve randomized neural networks for solving partial differential equations. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant PDE solver method, not language model evaluation or benchmark methodology.
Adaptive Randomized Neural Networks for PDEs: Structure-Aware Feature Redistribution and Residual-Driven Block Enrichment The paper introduces a two-stage adaptive feature construction method to improve randomized neural networks for solving partial differential equations. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Numerical PDE solution technique, not deployment safety or robustness in LLMs.
StochLOB: Heavy-Tailed Score Matching for Robust Limit Order Book Trend Prediction The paper proposes a heavy-tailed score-matching framework for limit order book trend prediction without needing reverse diffusion sampling at inference. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Limit order book forecasting, not AI safety, alignment, or guardrail systems.
StochLOB: Heavy-Tailed Score Matching for Robust Limit Order Book Trend Prediction The paper proposes a heavy-tailed score-matching framework for limit order book trend prediction without needing reverse diffusion sampling at inference. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Financial time series modeling paper, not LLM evaluation or benchmark design.
StochLOB: Heavy-Tailed Score Matching for Robust Limit Order Book Trend Prediction The paper proposes a heavy-tailed score-matching framework for limit order book trend prediction without needing reverse diffusion sampling at inference. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Diffusion modeling for finance, not small language models or efficient inference.
Noise-Space Reward Alignment for Few-Step Generative Models The paper proposes Noise Policy Optimization to align few-step generative models with black-box rewards by steering their initial input noise space. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Focuses on reward alignment in generative models, not deployment safety or jailbreak prevention.
Noise-Space Reward Alignment for Few-Step Generative Models The paper proposes Noise Policy Optimization to align few-step generative models with black-box rewards by steering their initial input noise space. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Presents reward alignment for generative models, not benchmark dataset or evaluation analysis.
Predicting Live Birth After Fresh Embryo Transfer in IVF/ICSI: Differential ICM/Trophectoderm Morphology Encoding and a Hybrid Attention--Logistic Regression Ensemble The paper presents an attention-based neural network and logistic regression ensemble to predict live birth outcomes before fresh embryo transfer. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Clinical tabular ensemble model, not small language models or inference serving.
Predicting Live Birth After Fresh Embryo Transfer in IVF/ICSI: Differential ICM/Trophectoderm Morphology Encoding and a Hybrid Attention--Logistic Regression Ensemble The paper presents an attention-based neural network and logistic regression ensemble to predict live birth outcomes before fresh embryo transfer. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Tabular clinical prediction for IVF, not LLM benchmarking or evaluation tools.
Predicting Live Birth After Fresh Embryo Transfer in IVF/ICSI: Differential ICM/Trophectoderm Morphology Encoding and a Hybrid Attention--Logistic Regression Ensemble The paper presents an attention-based neural network and logistic regression ensemble to predict live birth outcomes before fresh embryo transfer. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant IVF clinical outcome prediction, not LLM safety, guardrails, or red-teaming.
Pathwise Shrinkage and Probabilistic Calibration of the Multiplicative Half-Cauchy Process The paper analyzes theoretical properties of the multiplicative half-Cauchy process and provides a calibration rule linking prior beliefs to its scale parameter. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Pure Bayesian statistical theory, not model compression or inference efficiency.
Pathwise Shrinkage and Probabilistic Calibration of the Multiplicative Half-Cauchy Process The paper analyzes theoretical properties of the multiplicative half-Cauchy process and provides a calibration rule linking prior beliefs to its scale parameter. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Theoretical statistics paper on stochastic processes, not LLM evaluation.
Pathwise Shrinkage and Probabilistic Calibration of the Multiplicative Half-Cauchy Process The paper analyzes theoretical properties of the multiplicative half-Cauchy process and provides a calibration rule linking prior beliefs to its scale parameter. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Mathematical stochastic process analysis, not AI deployment safety or jailbreak defense.
Coordination-Free Workload Regularization for Variable-Cost Distributed Training The paper proposes a rank-local data-loading policy that rebalances variable-cost training batches without cross-node coordination to speed up distributed training. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Hardware workload balancing, not AI deployment safety, alignment, or guardrails.
Coordination-Free Workload Regularization for Variable-Cost Distributed Training The paper proposes a rank-local data-loading policy that rebalances variable-cost training batches without cross-node coordination to speed up distributed training. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on distributed training systems, not LLM evaluation benchmarks or metrics.
Sensor Diagnosis for Fuel and Emissions Monitoring in Heavy-Duty Diesel Vehicles The paper evaluates two machine learning procedures to diagnose faulty sensors and correct fuel and emission estimates in heavy-duty diesel engines. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Automotive engine sensor monitoring, not LLM safety, red-teaming, or guardrails.
Sensor Diagnosis for Fuel and Emissions Monitoring in Heavy-Duty Diesel Vehicles The paper evaluates two machine learning procedures to diagnose faulty sensors and correct fuel and emission estimates in heavy-duty diesel engines. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Deals with diesel sensor diagnostics, not LLM benchmark evaluation or judge testing.
Sensor Diagnosis for Fuel and Emissions Monitoring in Heavy-Duty Diesel Vehicles The paper evaluates two machine learning procedures to diagnose faulty sensors and correct fuel and emission estimates in heavy-duty diesel engines. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Automotive sensor fault diagnosis, not small language models or efficient inference.
Resisting Harmful Instructions through Training-Free Weight Editing The paper introduces a training-free weight-editing method to reduce large language model compliance with harmful instructions such as lying and sexism. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Presents a model editing technique rather than a benchmark or evaluation methodology.
Resisting Harmful Instructions through Training-Free Weight Editing The paper introduces a training-free weight-editing method to reduce large language model compliance with harmful instructions such as lying and sexism. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Focuses on alignment editing, not quantization, pruning, or efficient inference serving.
Poisoning the corpora of protein design systems during pre-training and retrieval The paper investigates data poisoning attacks targeting pre-training and retrieval corpora of protein language models and generative design systems. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Deals with protein generator security, not small model compression or inference efficiency.
Poisoning the corpora of protein design systems during pre-training and retrieval The paper investigates data poisoning attacks targeting pre-training and retrieval corpora of protein language models and generative design systems. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on protein model data poisoning rather than LLM benchmarking or evaluation.
Hard Way, The Only Way: Provable Merits of Training on Hardest Examples The paper provides a high-dimensional theoretical analysis of data pruning strategies when training models on the hardest examples. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Studies sample pruning theory in high dimensions, not model compression or serving.
Hard Way, The Only Way: Provable Merits of Training on Hardest Examples The paper provides a high-dimensional theoretical analysis of data pruning strategies when training models on the hardest examples. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Mathematical sample pruning analysis, not AI safety guardrails or deployment risks.
Hard Way, The Only Way: Provable Merits of Training on Hardest Examples The paper provides a high-dimensional theoretical analysis of data pruning strategies when training models on the hardest examples. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Theoretical analysis of data pruning rates, not LLM benchmark evaluation.
Learning and Forecasting Multiscale Spatiotemporal Epidemic Dynamics with Lagrangian-Eulerian Graph Neural ODEs The authors introduce a graph neural ordinary differential equation framework to forecast multiscale spatiotemporal epidemic spreading dynamics. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Focuses on spatio-temporal epidemic modeling, not LLM deployment safety or guardrails.
Learning and Forecasting Multiscale Spatiotemporal Epidemic Dynamics with Lagrangian-Eulerian Graph Neural ODEs The authors introduce a graph neural ordinary differential equation framework to forecast multiscale spatiotemporal epidemic spreading dynamics. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Epidemiological neural ODE architecture, not small or efficient language model serving.
Learning and Forecasting Multiscale Spatiotemporal Epidemic Dynamics with Lagrangian-Eulerian Graph Neural ODEs The authors introduce a graph neural ordinary differential equation framework to forecast multiscale spatiotemporal epidemic spreading dynamics. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Addresses epidemic dynamic forecasting models, not LLM benchmarks or evaluation methods.
Not All Change Is Damage: Update-Matched Unlearning for Large Language Models The paper introduces an update-matched evaluation reference to balance unlearning effectiveness with downstream learning ability in language models. Preprint	Small and efficient models via arxiv · 17 Oct, 06:18	Not relevant Investigates knowledge unlearning rather than model quantization, distillation, or serving speed.
Not All Change Is Damage: Update-Matched Unlearning for Large Language Models The paper introduces an update-matched evaluation reference to balance unlearning effectiveness with downstream learning ability in language models. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:18	Not relevant Focuses on machine unlearning methods rather than LLM benchmarking or judge reliability.
Compatibility Matters: A Framework for Token-Level Teacher Selection in On-Policy Distillation The authors propose a theoretical and empirical framework to route token-level teacher selection during on-policy student distillation. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:18	Not relevant Centers on distillation policy routing, not deployment safety or alignment robustness.

3,781–3,840 of 17,313← Previous Next →