TechniqueRLHF / Alignment8 recent entries12 Oct 2026Evaluating LLM Creativity through Long-Tail PerformancearXiv:2610.12998v1 Announce Type: new Abstract: Creativity benchmarks for large language models (LLMs) vary widely in task and metric design. This raises the question of whether they capture a common →12 Oct 2026EmoGavel: Grounding-Aware Reinforcement Learning with Verifiable Rewards for Multimodal Emotion Preference LearningarXiv:2610.13353v1 Announce Type: cross Abstract: Multimodal emotion preference learning uses pairwise comparisons to identify which open-ended emotion descriptions align more closely with human judgm
TechniqueRAG8 recent entries12 Oct 2026Locate, Judge, Read: Role-Specialized Tool Agents over a Knowledge-Point Index for Open-Domain Multimodal Document QAarXiv:2610.13128v1 Announce Type: new Abstract: Question answering over large heterogeneous document collections requires first finding the right passages and then reading them, yet current retrieval-→12 Oct 2026Knowing Is Not Telling: Identity Ossification in LLM Representations of Trans-border CommunitiesarXiv:2610.12610v1 Announce Type: cross Abstract: Ethnic and cultural identities rarely align with political borders, a tension made sharper by migration and long-standing cross-border ties. Communiti→12 Oct 2026HART: Hybrid Attention-RWKV Time series Foundation ModelarXiv:2610.12883v1 Announce Type: new Abstract: Non-autoregressive multi-step forecasting (NAF) enables time series foundation models to predict an entire future horizon in a single forward pass. Desp→12 Oct 2026GeopoliticalQA: Benchmarking Retrieval and Generation over Intelligence ReportsarXiv:2610.12617v1 Announce Type: new Abstract: Geopolitical and strategic intelligence reports constitute a rich source of expert knowledge and have increasing value in the current political climate,→12 Oct 2026DRYLAB-Bench: Benchmarking the Biosecurity Risk of Agentic LLM Mutation DesignarXiv:2610.13116v1 Announce Type: new Abstract: As large language models (LLMs) acquire advanced biological design capabilities, they offer substantial research benefits but simultaneously introduce s→12 Oct 2026Can Language Models Play the Wikipedia Game?arXiv:2610.13948v1 Announce Type: new Abstract: Language models increasingly support or automate information retrieval from document collections too large to index. This task is difficult when it requ→12 Oct 2026AgentTune: A Unified Training Lifecycle for Tool-Using LLM AgentsarXiv:2610.13335v1 Announce Type: new Abstract: Tool-using agents increasingly require a common substrate for building, training, evaluating, distilling, and maintaining agentic systems, yet existing →12 Oct 2026A 'free' multi-modal 'embedding' model (already) included with your LLM: vLLM plugin + upstream PRDisclaimer: I work for NuMind, the company behind this plugin. TLDR: Text/image (simili-)embeddings from the LLM you already have loaded :-) I've been working on making it easier to obtain embeddings
TechniqueAgents8 recent entries12 Oct 2026BT-Agent: Deploying Onboard Robot Agents with Behavior Trees and Multi-Fidelity EvaluationarXiv:2610.13862v1 Announce Type: new Abstract: Deploying language-model agents onboard mobile robots requires task reasoning and responsive execution within limited compute and memory. We propose BT-→12 Oct 2026BrokenML: can agents fix broken ML jobs?Write-up: https://zhebrak.io/posts/brokenml/ Repo: https://github.com/zhebrak/brokenml I built a benchmark to determine the access level agents need to successfully fix broken ML workloads (PyTorch ov→12 Oct 2026Bistable by Construction: Wall-Clock-Calibrated State Monitors Have No Moment-Detection Regime at Agent CadencearXiv:2606.19386v2 Announce Type: replace-cross Abstract: Runtime monitors for autonomous agents commonly threshold an accumulated internal state - a behavioural baseline, a drift statistic, or, in ou→12 Oct 2026Beyond Temporal Proximity: Explanatory World Models for Credit Assignment in Reinforcement LearningarXiv:2610.13504v1 Announce Type: cross Abstract: Generative Recommenders increasingly use Reinforcement Learning (RL) to optimize outcomes that may appear hundreds or thousands of interactions after →12 Oct 2026APEX-Harm: Evaluating Agent Safety in Shared Professional WorkspacesarXiv:2610.13623v1 Announce Type: cross Abstract: Professional agents increasingly work through command-line interfaces and multiple tools over shared workspaces containing files, messages, and busine→12 Oct 2026AgentTune: A Unified Training Lifecycle for Tool-Using LLM AgentsarXiv:2610.13335v1 Announce Type: new Abstract: Tool-using agents increasingly require a common substrate for building, training, evaluating, distilling, and maintaining agentic systems, yet existing →12 Oct 2026AgentSWE: Can Coding Agents Build the Agent You Actually Want?arXiv:2610.13361v1 Announce Type: cross Abstract: As the production of software passes into the hands of software itself, coding agents are increasingly commissioned to build agents fitted to their us→12 Oct 2026A Baseline Assessment for the Agentic Web Technology StackarXiv:2610.13703v1 Announce Type: new Abstract: The fast pace in which protocols, frameworks, and technologies are being proposed for LLM-based agents results in a complex scenario for anyone aiming t
TechniqueFine-tuning8 recent entries12 Oct 2026Hindsight Structured Advice Distillation: Steering Executor Agents with Advisors Trained on Their Failures and Task CharacteristicsarXiv:2610.13000v1 Announce Type: new Abstract: Language-model agents are used for specialized work such as financial research, healthcare and law. Such agents are often capable enough for a task but →12 Oct 2026Expert-level detection of low-voltage intervals in EEG as a robust biomarker of neonatal encephalopathyarXiv:2610.12680v1 Announce Type: new Abstract: Background: Discontinuity in neonatal EEG is a critical marker of brain dysfunction but can be difficult to quantify visually. Inter-burst intervals (IB→12 Oct 2026EmoGavel: Grounding-Aware Reinforcement Learning with Verifiable Rewards for Multimodal Emotion Preference LearningarXiv:2610.13353v1 Announce Type: cross Abstract: Multimodal emotion preference learning uses pairwise comparisons to identify which open-ended emotion descriptions align more closely with human judgm→12 Oct 2026Cross-Behavior Policy Optimization: Unified Learning from Offline Demonstrations and On- and Off-Policy Rollouts via a Shared Regression TargetarXiv:2610.12830v1 Announce Type: new Abstract: Reasoning post-training draws on offline demonstrations and policy-generated rollouts, yet conventional supervised fine-tuning (SFT) followed by reinfor→12 Oct 2026Byzantine-Resilient Federated Large Language Model via Hypergraph Message PassingarXiv:2610.12711v1 Announce Type: new Abstract: Federated Large Language Models (FedLLMs) are promising to deliver privacy-preserving, collaborative intelligence. The decentralized training process is→12 Oct 2026BLEND: Balancing Personalization vs. Generalization in Federated Vision-Language ModelsarXiv:2610.12907v1 Announce Type: cross Abstract: Federated vision-language models (FedVLMs) represent one of the emerging frontiers of machine learning, aiming to integrate federated learning within →12 Oct 2026Behind-the-Meter Photovoltaic Visibility from Net Smart Meters: Deployment Regimes and Operational InsightsarXiv:2610.12527v1 Announce Type: new Abstract: Distribution operators need visibility of behind-the-meter photovoltaic (PV) generation, yet most residential customers are observed only through net sm→12 Oct 2026A Categorical Atlas of Parameter-Efficient Fine-Tuning: Cones, Orbits, and GaugesarXiv:2610.12728v1 Announce Type: new Abstract: Parameter-efficient fine-tuning has hundreds of methods, usually compared by parameter count and benchmark score. We compare them as maps. A pretrained
TechniqueMultimodal8 recent entries12 Oct 2026Decision 3.0: open, multimodal decision models from 0.6B to 27B. Ranks #1 on both Jev Decision Index boardsWith the vLLM Semantic Router team we released Decision 3.0 (d3), our new SOTA decision model family ranging from 0.6B to 27B, Apache-2.0, all reading text, images and video. How it works - Built on Q→12 Oct 2026DEAF: A Benchmark for Diagnostic Evaluation of Acoustic Faithfulness in Audio Language ModelsarXiv:2603.18048v3 Announce Type: replace Abstract: Recent Audio Multimodal Large Language Models (Audio MLLMs) demonstrate impressive performance on speech benchmarks, yet it remains unclear whether →12 Oct 2026CORE-Seg: Reasoning-Driven Segmentation for Complex Lesions via Reinforcement LearningarXiv:2603.05911v2 Announce Type: replace-cross Abstract: Medical image segmentation is undergoing a paradigm shift from conventional visual pattern matching to cognitive reasoning analysis. Although →12 Oct 2026Can Language Models Play the Wikipedia Game?arXiv:2610.13948v1 Announce Type: new Abstract: Language models increasingly support or automate information retrieval from document collections too large to index. This task is difficult when it requ→12 Oct 2026BLEND: Balancing Personalization vs. Generalization in Federated Vision-Language ModelsarXiv:2610.12907v1 Announce Type: cross Abstract: Federated vision-language models (FedVLMs) represent one of the emerging frontiers of machine learning, aiming to integrate federated learning within →12 Oct 2026AudioTETRIS: Token Elimination via Transitional Importance Signals for Efficient Audio LLMsarXiv:2610.13600v1 Announce Type: cross Abstract: Large Audio Language Models (LALMs) achieve strong performance on speech-text tasks, but standard audio encoders produce hundreds of tokens for even s→12 Oct 2026AdaC-GRPO: Towards Mastering Long-Horizon Visual Sudoku Puzzles via Adaptive Curriculum LearningarXiv:2610.13184v1 Announce Type: new Abstract: While reinforcement learning with verifiable reward (RLVR) enhances reasoning in large language models, its efficacy in long-horizon multimodal tasks re→12 Oct 2026A 'free' multi-modal 'embedding' model (already) included with your LLM: vLLM plugin + upstream PRDisclaimer: I work for NuMind, the company behind this plugin. TLDR: Text/image (simili-)embeddings from the LLM you already have loaded :-) I've been working on making it easier to obtain embeddings
TechniqueSafety8 recent entries12 Oct 2026Scalable Circuit Learning for Interpreting Large Language ModelsarXiv:2606.16939v2 Announce Type: replace-cross Abstract: A prominent research direction in mechanistic interpretability is learning sparse circuits over LLM components to reveal how they jointly prod→12 Oct 2026REFLECT: Role-Aware Detection of Sentiment, Emotion, Politeness, and FormalityarXiv:2610.13797v1 Announce Type: new Abstract: Detecting communicative meaning in dialogue requires modeling not only lexical content but also the participant-role structures through which sentiment,→12 Oct 2026ReasoningShield: Safety Detection over Reasoning Traces of Large Reasoning ModelsarXiv:2505.17244v3 Announce Type: replace-cross Abstract: Large Reasoning Models (LRMs) leverage explicit reasoning traces, known as Chain-of-Thought (CoT) reasoning, to decompose complex problems int→12 Oct 2026FORTE: A Production Framework for Feature Evaluation and Deprecation in Industrial Recommender SystemsarXiv:2610.13806v1 Announce Type: cross Abstract: Production recommender systems have feature-lifecycle asymmetry: adding new features must produce gains on some metric, while feature removal must not→12 Oct 2026CoBe: A Benchmark For Conversational Counterfactual Text EditingarXiv:2610.13057v1 Announce Type: cross Abstract: Counterfactual text editing is the ability to revise a given text in light of a hypothetical change while preserving facts not causally affected by th→12 Oct 2026Bistable by Construction: Wall-Clock-Calibrated State Monitors Have No Moment-Detection Regime at Agent CadencearXiv:2606.19386v2 Announce Type: replace-cross Abstract: Runtime monitors for autonomous agents commonly threshold an accumulated internal state - a behavioural baseline, a drift statistic, or, in ou→12 Oct 2026Attention Sees, MLP Refuses: Tracing the Attention-to-MLP Safety Pathway in LLMsarXiv:2610.13078v1 Announce Type: new Abstract: Safety-aligned large language models (LLMs) remain vulnerable to jailbreaks and can over-refuse benign requests. Understanding these failures requires d→12 Oct 2026APEX-Harm: Evaluating Agent Safety in Shared Professional WorkspacesarXiv:2610.13623v1 Announce Type: cross Abstract: Professional agents increasingly work through command-line interfaces and multiple tools over shared workspaces containing files, messages, and busine