Every paper checked — nothing hidden

All papers

Every paper checked in the week of 12–18 Oct 2026 · Small and efficient models — nothing is hidden. Brief papers first.

All2,430 In brief3 More (on-topic)455 Not relevant1,957 Not assessed6 Older27 Could not confirm6 Blogs with no paper16 All blogs19

Paper	Topic	Relevance
RealityBench: Simulated Environments to Evaluate a Vision Language Model's Ability to Document Human Physical Work The paper presents a simulated benchmark across eleven unconstrained environments to evaluate vision-language models on documenting real-world human physical work. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not investigate model compression, quantization, or inference acceleration.
Memory Without Curation: Learning Multi-Stream Latent Memory for Robot Policies The paper proposes an uncurated multi-stream latent memory framework that augments vision-language-action robotic policies while maintaining low inference latency. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Develops a latent memory framework for robots rather than general model compression.
Target-ESS Tempered Policy Optimization (TETPO): Smooth Variance Control via Bridge Importance Sampling The paper proposes a data-adaptive importance sampling method to stabilize variance and prevent sample size collapse in asynchronous RL post-training. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on RL algorithm stability rather than model compression or inference efficiency.
Jevolution: A Framework for Studying Communication and Collective Dynamics in AI Agents The paper presents an instrumented predator-prey simulation framework to evaluate communication, scheduling, and collective dynamics in remote decision-making agents. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model quantization, distillation, or efficient inference.
Diffusion-Based Auctions for Multi-Winner Banner Ads The paper proposes a multi-winner banner ad auction mechanism using diffusion models guided by bids and text descriptions to generate ads. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model efficiency, distillation, or lightweight architectures.
RES-MR: Risk-Aware Reasoning for Explainable and Safe Medication Recommendation The paper introduces a two-stage reasoning framework to generate personalized, safe, and explainable medication recommendations using language models. Peer-reviewed	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not target efficient inference, quantization, or small on-device models.
Scale, Drivers, and Limits of Representational Alignment Across Large Language Models The paper presents two geometric methods to evaluate where and why representational structures align across different large language models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, quantization, distillation, or serving efficiency.
AphaVoice: Text-to-Speech Model for Aphasia Patient Simulation The paper develops a text-to-speech model that simulates realistic clinical dysfluency in aphasia speech across multiple speakers for training healthcare professionals. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, pruning, or efficient inference.
Active Neural Matter: A Finite-Window Response Framework for Physical Systems and Distributed AI Workflows The paper introduces a finite-window theoretical framework to determine what historical state information dynamical systems and distributed AI workflows must retain. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Deals with abstract dynamical systems rather than model compression or efficient serving.
Decoupling Knowledge Forgetting and Refusal Learning in MLLM Unlearning The paper proposes a distillation framework to remove specific multimodal knowledge without causing hallucinations or leaving residual queryable information. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on knowledge erasure rather than model compression or inference efficiency.
EarthAlign: Learning Language-Aligned Representations of Earth System Data The paper releases a multimodal dataset and dual-encoder model to align natural language descriptions with global Earth system meteorological data. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not focus on model compression, quantization, or efficient serving.
Towards Merging-conditional Backdoor Unalignment against Large Language Models This paper demonstrates a security risk where dormant backdoors trigger safety failures only after merging models that passed individual safety audits. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies backdoor vulnerabilities in model merging, not efficiency, pruning, or serving.
How to Build a Benchmark That Actually Measures Taste This paper introduces a pairwise evaluation benchmark and protocol designed to measure subjective aesthetic taste while controlling for shortcut artifacts. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model distillation, quantization, pruning, or efficient inference.
Vela: User Behavior as Supervision Sparks LLM Intelligence This paper presents a framework that converts long-term real user behavior records into supervisory tasks to train personalization-focused language models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on behavior supervision data, not model compression or efficiency.
Integrating AIRL-Based Reward Estimation into a Control Method Using Fuzzy Inference and Policy Gradient This paper integrates adversarial inverse reinforcement learning into a fuzzy control framework to estimate reward functions directly from demonstration data. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on fuzzy RL control, not model efficiency or distillation.
Self-improving data synthesis can scale embedding models across domains and tasks This paper introduces a self-improving synthetic data generation pipeline to train embedding models across diverse retrieval domains without extensive manual annotation. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on training data synthesis rather than model compression or efficient serving.
PTGPO: Prefix Tree Group Policy Optimization This paper proposes a critic-free reinforcement learning algorithm using prefix trees to assign step-level credit for interactive multi-step tasks. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on RL credit assignment rather than model compression or inference efficiency.
CUSE: Fine-Grained Multimodal Unlearning with Counterfactual Guidance and On-Policy Self-Distillation This paper introduces a multimodal unlearning method using counterfactual inputs and on-policy self-distillation to remove sensitive data without damaging general knowledge. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Covers selective forgetting, not small models, quantization, or inference acceleration.
Prior-Fitted and Frozen Scorers for Multimodal Contrastive Learning This paper investigates using frozen, pre-trained Transformer scorers instead of fixed mathematical formulas for multimodal contrastive representation learning. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on frozen contrastive scorers, not model compression or efficient serving.
CESNET-TimeSeries26: A Multi-Year Network Traffic Time Series Dataset for Annual Seasonality, Concept Drift, and Long-Term Anomaly Detection This paper presents a multi-year network traffic time-series dataset to evaluate long-term anomaly detection and forecasting models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies network data forecasting, not efficient model compression or serving.
MARL-BattleGrounds: A Novel JAX-native Benchmark for Heterogeneous and Competitive Multi-agent RL This paper introduces a JAX-native benchmark for evaluating heterogeneous and competitive multi-agent reinforcement learning algorithms. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Benchmark compute efficiency in JAX, not small LLM compression or serving.
Bridging Risk Approximation Gaps in Model Predictive Task Sampling via In-Context Modeling This paper addresses failure modes in model predictive task sampling for reinforcement learning by using in-context modeling. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Addresses RL task sampling efficiency, not model compression or lightweight serving.
WorkBorne: Discovering Agentic Tasks from Real-World Artifact Transitions This paper presents a framework for discovering and synthesizing terminal-based agent tasks grounded in real-world artifact transitions. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, quantization, or inference efficiency.
SkillForgeBench: Evaluating Agent Reliability Grounded in Agent Skill Specifications This paper introduces a benchmark that uses agent skill specifications to evaluate the reliability and procedural correctness of AI agents. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not focus on model efficiency, distillation, or quantization.
Predicting how model propensities generalize out of distribution This paper studies how language model behavioral propensities generalize out-of-distribution across different elicitation methods. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, quantization, distillation, or serving efficiency.
Probabilistic Multi-Source Data Selection via Learnable Categorical Priors: Variance Reduction and Guaranteed Derandomization This paper introduces a probabilistic relaxation method with learnable categorical priors for bilevel multi-source data selection. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on dataset selection algorithms rather than model compression or inference efficiency.
Counterfactual Rollouts for Offline Contrastive Goal-Conditioned Reinforcement Learning This paper proposes a causal contrastive reinforcement learning method using counterfactual rollouts to handle unobserved confounding in offline demonstrations. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Deals with RL counterfactual transitions rather than model compression or efficient serving.
What Survives When Activation Steering Is Used as Memory This paper evaluates the capabilities and limitations of using activation steering vectors as an agent memory mechanism. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Analyzes vector memory retention rather than model compression or inference acceleration.
AlphaRLM: Aligning Language Models for Literature-Grounded Symbolic Factor Discovery This paper trains language models using reinforcement learning to construct executable financial trading formulas grounded in research literature. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on RL formulation for factor mining rather than model compression or efficient inference.
Fitting Every Hint Is Not Enough: Premature Floods in Recurrent Algorithmic Reasoners This paper investigates failure modes where recurrent algorithmic reasoning models jump ahead prematurely when generalizing to larger graphs. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies algorithmic reasoner generalization, not model compression or efficient serving.
Mixing Regularization for Average-Reward $Q$-Learning and $Q$-Planning This paper provides theoretical convergence guarantees for average-reward reinforcement learning algorithms via mixing policy regularization. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Deals with reinforcement learning optimization rates, not model efficiency or distillation.
InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly Detection This paper introduces a benchmark to evaluate continual learning in semi-supervised visual industrial anomaly detection across components. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on continual learning benchmarks rather than model compression or efficient serving.
Belief Embedding Tree Search This paper proposes a tree search method using belief embeddings for planning in partially observable environments. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on POMDP belief representation, not LLM compression or inference efficiency.
Select or Rewrite: Oracle-Guided Fine-Tuning for Generative ASR Post-Editing from N-Best Hypotheses This paper fine-tunes language models to selectively correct speech recognition transcripts using curated candidate pools. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies data supervision strategies rather than model pruning, distillation, or quantization.
Have You Been Convinced? A Neuro-symbolic Argumentation Reasoning Approach to Detecting Persuasiveness in ChangeMyView Discussions This paper presents a neuro-symbolic framework using language models and formal argumentation semantics to detect persuasiveness in online discussions. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model quantization, distillation, or efficient inference serving.
scHFBiMamba: a bidirectional Mamba autoencoder for cardiac disease-state classification from single-nucleus RNA-seq data Evaluates a compact bidirectional Mamba state-space autoencoder for classifying heart failure disease states from single-nucleus RNA sequencing data. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Describes a compact biological autoencoder rather than efficient LLM serving techniques.
To Defer or Not to Defer? Online Learning with LLMs and Congested Experts Proposes an online learning algorithm that decides whether language models should answer queries directly or defer to congested human experts. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model distillation, quantization, or efficient model architectures.
UDHRBench: Benchmarking Human Rights in Large Language Models Evaluates commercial language models against Universal Declaration of Human Rights principles, revealing misalignments driven by efficiency arguments. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Evaluates model ethics, not inference optimization or model compression.
SocraticRepair: Planning-Guided Socratic Tutoring for Program Debugging and Misconception Repair Fine-tunes a coding model using synthetic tutoring dialogues to provide step-by-step guidance without revealing program fixes prematurely. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Deals with Socratic tutoring strategies, not model compression or efficiency.
PANINI-DR: Controlled Deep Research through Generative Semantic Workspaces Presents a multi-agent framework that decouples research planning, subquestion investigation, and factual verification via a shared semantic workspace. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Deals with research workflows, not small models or quantization techniques.
EvolvAttack: Backdooring self Evolving VLM Agents The paper introduces EvolvAttack, demonstrating a backdoor mechanism that activates only after a vision-language agent evolves its memory over time. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on backdoor vulnerabilities rather than model compression or inference efficiency.
Sequent-Prover: Training Agents for Formal, Checkable SMT-based Reasoning The paper introduces Sequent-Prover, training language model agents to perform multi-turn formal reasoning using an external SMT solver scratchpad. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on agent fine-tuning and reasoning accuracy, not efficiency techniques.
LeHami: Symplectic Latent Dynamics for World Models The paper proposes LeHami, incorporating symplectic Hamiltonian dynamics into latent predictive architectures to prevent representation collapse over long rollouts. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on symplectic dynamics preservation, not model compression or quantization.
North or South? Diagnosing LLMs' Understanding of North Korean Language, Society, and Knowledge The paper presents LENS, a benchmark evaluating language models on North Korean linguistic nuances, sociocultural perspectives, and domain knowledge. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Presents an evaluation benchmark rather than methods for model efficiency.
Protein Crystallization Segmentation and Classification Using Subordinate Color Channel in Fluorescence Microscopy Images The paper proposes using subordinate color channel intensities in fluorescence microscopy to improve protein crystallization segmentation and classification. Peer-reviewed	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Proposes color channel segmentation, not model compression or efficient serving.
WisdomNet: trustable machine learning toward error-free classification The paper proposes an artificial neural network framework designed to reject uncertain predictions to prevent misclassification errors. Peer-reviewed	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies selective classification, not model compression or efficient serving.
Accuracy and Uncertainty in Machine Deception Detection (Research White Paper) The paper proposes a multi-layer detector ensemble to identify deceptive model outputs while calibrating score uncertainty and false-positive rates. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies activation-based deception detection, not model efficiency or quantization.
SuperAI: Merging Capabilities Across Heterogeneous Foundation Models The paper introduces SuperAI, a framework to merge complementary capabilities across heterogeneous foundation models using low-dimensional subspace alignment. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Merges heterogeneous model representations rather than compressing models for serving.
Anchor-and-Harvest Curriculum: A Fully Unmasked Non-LLM Curriculum for Craftax The paper introduces a rule-based curriculum training system that improves reinforcement learning performance on Craftax without language models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies RL game curricula, not model quantization or serving efficiency.
ZOIF: Zeroth-Order Influence Function Estimation for LLM Data Selection Using Only Forward Passes The paper introduces a zeroth-order influence function framework to enable memory-efficient training data selection using only forward passes. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies zeroth-order data attribution rather than model compression or inference acceleration.
CodeWalk: Generating Coding Benchmarks by Walking a Problem Graph The paper presents CodeWalk, a method to create challenging coding benchmarks at low cost by programmatically composing verified atomic problems. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Discusses benchmark composition rather than model quantization or inference efficiency.
Siam-ALS: An Explainable Siamese Deep Learning Framework for Early ALS Detection Using EEG and Eye-Tracking Signals The paper introduces an explainable Siamese deep learning framework combining synchronized EEG and eye-tracking signals for early ALS detection. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, quantization, or efficient LLM serving.
Adversarial Detection for Pre-trained Classifiers via Lipschitz-Margin Analysis with Certification The paper proposes certified adversarial detection using Lipschitz-margin analysis to detect perturbations rather than forcing invariant predictions. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Presents theoretical robustness bounds without focusing on compression or serving efficiency.
TRACE: Temporal Risk Assessment and Concept Editing for Safer Language Generation The paper introduces TRACE to monitor harmful response risks during generation, edit token activations selectively, and maintain auditable decision traces. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on risk mitigation rather than model compression or inference efficiency.
A Sovereign, Indic-First, Reasoning Model The paper presents a hybrid language model adapted for reasoning, instruction following, and tool use across English and eight Indian languages. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Deals with a 30B MoE pretraining rather than efficient model compression.
In-Place Indic Tokenizer Expansion and Continual Pretraining of Nemotron-3-Nano-30B-A3B This paper details vocabulary expansion and continual pretraining techniques to adapt an open language model across eight Indic languages. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Addresses multilingual pretraining rather than model compression or inference efficiency.
Sensitivity or Criterion? Measuring and Moving When Language Models Act This paper introduces a diagnostic framework using signal detection theory to measure when language model agents decide to use tools. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Analyzes model sensitivity and criterion, not compression or efficiency.
Pessimistic Latent Task-aware Optimization for Robust Offline Meta-Reinforcement Learning This paper proposes a pessimistic offline meta-reinforcement learning method that generates synthetic off-support latents to improve out-of-distribution task transfer. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on RL latent representations, not model compression.
World Modeling through Spectral Alignment This paper introduces a world-modeling framework that uses kernel similarity to control which physical properties are preserved in learned embeddings. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Addresses representation geometry, not model compression or efficient serving.
Hidden in Plain Sight: Advancing Harmful Content Understanding via Perception-Aware Adversarial Self-Play This paper presents an adversarial self-play framework using multi-perceptual views to improve language model detection of obfuscated harmful text. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not focus on model compression, quantization, or distillation.

1,741–1,800 of 2,430← Previous Next →