标题：Same Luck, Different Choice: Twin-World Credit Assignment for Agentic Reinforcement Learning
来源：arXiv
链接：https://arxiv.org/abs/2610.31544
摘要：Agentic reinforcement learning rewards a complete episode, yet the policy needs credit for every decision inside it. Counterfactual branching estimates this credit by re-running the episode after an alternative decision, but the branch draws its sampling noise independently of the factual episode, so the outcome difference mixes the effect of the decision with the luck of everything that follows. We observe that in agentic search the retriever, the corpus, and the grader are deterministic, so an episode is a deterministic function of the policy's decisions and its sampling noise. Twin worlds are therefore realizable: the same noise, a different decision. We introduce TWICE (TWIn-world Credit Estimation), which keys the Gumbel noise of every token by turn and line, rolls out counterfactual
标题：Is Complex Agentic Memory Always Needed? Naive RAG Is Stronger Than We Thought
来源：arXiv
链接：https://arxiv.org/abs/2610.30755
摘要：LLM memory is widely used in long-term conversation understanding. Agentic memory systems often use mutation, structural organization, learned retrieval strategies, or other complex operations. Is this complexity always necessary? We study Raw-RAG, a naive RAG baseline using dense retrieval and reranking over raw history to supply evidence to a reader LLM. Across LoCoMo, LongMemEval, MEME, and MemoryAgentBench, Raw-RAG ranks first among the compared memory systems in six of seven benchmark-model settings, outperforming the strongest compared agentic memory system by 0.13-5.53 points. We investigate why raw-history retrieval remains competitive. Evidence analysis shows that its higher coverage of necessary evidence can offset lower utilization. Useful evidence includes both answer values an
标题：OpenAttention: One Percent Is Enough with Tail Proxies
来源：arXiv
链接：https://arxiv.org/abs/2610.30951
摘要：Sparse attention reads only a small subset of the key-value cache, and how to treat the unread tail remains a key challenge. Discarding the tail biases the output, and sampling it as previous approaches do fails at reads of 5% and below, where the tail likely holds heavy hitters the selector missed. Our key observation is that the simplest possible proxy, one token that folds the entire tail into its count-aware mean, already yields lower output error than a sample at every read behind a practical selector. We further show in closed form that at low reads sampling nearly doubles the weight error of discarding while a single proxy stays below it. Building on these findings, we explore approximating the dense attention distribution with a few proxies that target the tail. We propose OpenAtte
标题：Conservative Meta-Agent Search: Architecture Adaptation under Non-Stationarity
来源：arXiv
链接：https://arxiv.org/abs/2610.31436
摘要：Automated agentic-system design is typically evaluated on static task suites with immediate scalar rewards. Deployed systems face persistent shift, delayed dependent feedback, correlated component failures, and costly architecture changes. We formalize meta-agent architecture search as sequential incumbent replacement with switching costs and hard risk constraints and introduce Conservative Meta-Agent Search (CMAS). CMAS retrieves archived architectures and proposes typed graph edits, then scores each against the incumbent by paired shadow replay under common random numbers, with interventional component credit. Promotion requires an empirically calibrated, variance-penalized simultaneous score above a pre-specified margin, and is followed by a simulated canary stage and pre-specified roll
标题：AgenticEvolve: Evolutionary Program Search with Persistent Language Model Agents that Design and Run Experiments
来源：arXiv
链接：https://arxiv.org/abs/2610.31326
摘要：Evolutionary program search with large language models keeps the structure of an evolutionary algorithm and uses the model as its variation operator. Typically, the model generates each candidate program in a single request from parent programs and their scores. We present AgenticEvolve, an evolutionary program search framework in which each island is run by a persistent large language model agent. Between evaluations, the agent designs and runs experiments on the candidate program. To measure the contribution of experimentation, we compare AgenticEvolve with No-analysis, an otherwise identical configuration with experiments disabled. On two heuristic design tasks with three large language models, AgenticEvolve produces better programs than No-analysis and OpenEvolve on both hidden and uns
标题：Agentic Knowledge-Model Co-Evolution via Consistency Regularization
来源：arXiv
链接：https://arxiv.org/abs/2610.30790
摘要：LLM agents can evolve by learning from accumulated experience. Existing approaches extract reusable external knowledge, such as skills and memories, from experience to guide subsequent task solving, with some further internalizing this knowledge through model training. However, different ways of analyzing and transforming the same experience can yield knowledge with substantially different utility, even when the resulting guidance appears reasonable. Inspired by consistency regularization, we identify consistent utility gains across different ways of analyzing the same experience, rather than relying on any single way of analyzing it. We introduce SPLICE, a framework for consistency-regularized knowledge-model co-evolution. It constructs evolution views by perturbing how the same experienc
标题：AfterDeleteBench: Auditing Post-Deletion Information Persistence in LLM Agents
来源：arXiv
链接：https://arxiv.org/abs/2610.31366
摘要：Many LLM agents maintain external persistent memory and derive additional state from past interactions, such as summaries. Deleting an original memory record therefore need not remove information carried by its descendants. We introduce AfterDeleteBench, a controlled behavioral audit for this failure mode in fixed-weight LLM agents. For each synthetic secret, we compare an exposed-then-deleted agent with an otherwise matched agent that was never exposed to the secret, allowing post-deletion recovery to be attributed to the original exposure. In our primary setting, a single query recovers the deleted secret from Qwen2.5-7B-Instruct in roughly two thirds of cases, while matched never-exposed controls have zero recovery. Crucially, recovery does not require a surviving verbatim copy: when th
OpenAI 发布 Decisions API，面向低延迟低成本决策
OpenAI 正式开放基于 GPT-6 Luna 的 Decisions API，以低延迟、低成本及多模态批处理能力支持分类、路由、评分等答案范围明确的实时决策任务。
标题：ADAPT for Science: A Multi-Agent System for Self-Evolving Scientific Optimization
来源：arXiv
链接：https://arxiv.org/abs/2610.30851
摘要：Scientific discovery often requires identifying a small number of high-performing designs in vast, constrained search spaces. This is difficult when evaluations are costly, feasible solutions are sparse, parameters are strongly coupled, and physical constraints are strict. Recent AI for AI methods can automatically improve algorithms and optimization procedures. However, general purpose approaches are not designed around scientific priors, physical constraints, and limited evaluation budgets. We introduce ADAPT for Science, a multi-agent framework that treats scientific optimization as two coupled decisions: where to search and how to search. ADAPT uses historical designs to identify promising and diverse initial conditions, progressively concentrates computation on favorable search direct
标题：What the Specification Did Not Say: Specification-Oracle Gaps in Test-Graded Coding-Agent Benchmarks
来源：arXiv
链接：https://arxiv.org/abs/2610.30635
摘要：Coding-agent benchmarks often describe a task in natural language but grade it with tests taken from the upstream project. Passing those tests can require information the description never states: a constant's name, a module path, an exact message. We call this a specification-oracle gap: where it exists, a benchmark also measures whether an agent reproduces upstream implementation details. We formalise the gap, including the inconsistencies that chains of dependent tasks add, and build a static audit for five languages that classifies every name-, path- and literal-level dependency of every scored test (held-out precision 89.2%, recall 92.9%). On SWE-Milestone (98 graded milestones, 35 published runs), 25 milestones contain a test that depends on an unstated, non-inferable name or path. B
基于 Petri 网监督控制的智能体认证执行框架
论文将智能体的工具执行框架形式化为 Petri 网监督控制问题，在提示注入和不可控工具结果下实现安全且尽量少干预的动作执行。
论文在 LLM 推理层与工具执行层之间引入基于 Petri 网的认证执行框架，用标记表示能力、审批和预算，仅在前置条件满足时允许动作转换。系统将工具结果视为不可控事件，通过传统最大许可监督器实施安全约束，并在四项假设下将形式化安全保证迁移到实际运行系统。针对智能体遭拒后重新规划和重复提议的闭环，作者指出传统监督器可能比更严格的安全监督器产生任意程度的额外干预。对于给定提议模型，最少干预的最大许可监督器可表示为传统监督器的价值阈值细化，并能以至多一个监控库所约束并发动作顺序。自主发布智能体实验表明，安全部署需要保留在发布流程启动后拒绝某一步骤的能力。
用途与启示
为智能体工具调用提供与自然语言表述无关的能力、审批及预算约束，可抵御提示注入诱导的越权执行。
将执行安全转化为可综合、可验证的 Petri 网监督控制问题，为高风险智能体提供形式化安全保证。
揭示最大许可并不等同于重规划闭环中的最少干预，设计执行监控器时还需考虑拒绝后的推理成本与后续策略。
可用于软件发布、运维和其他具有不可逆动作的智能体系统，在任务启动后仍保留动态阻断能力。
📝 原始笔记（逐字保留）
标题：Certified Execution Harness for Agentic AI: Safe and Least-Intrusive Petri-Net Supervisory Control Between Reasoning and Action
来源：arXiv
链接：https://arxiv.org/abs/2610.31376
摘要：This paper studies the problem of safe tool execution for large language model (LLM) agents under prompt injection. Using Petri nets, we formalize the execution harness between the LLM and its tools as a supervisory control problem, in which marks represent capabilities, approvals and budgets. Since a transition fires only if the marks it needs are present, an injected request whose preconditions are unmet or already spent is refused regardless of its wording. Treating tool outcomes as uncontrollable, we enforce safety by the conventional maximally permissive supervisor, and this safety transfers to the running system under four assumptions. Moreover, to capture an agent that may propose again after a refusal, we introduce the notion of a replanning closed loop, in which each retry costs a
标题：Jev-DLC: Decision-Level Continual Learning from Typed Feedback
来源：arXiv
链接：https://arxiv.org/abs/2610.31521
摘要：Individual preferences often become visible only after deployment, when a user accepts one decision or revokes it in favor of another. Jev-style typed interfaces make this feedback explicit by identifying the selected action and, under revocation, the decision it replaces. We study how these consequences persist across requests as Decision-Level Continual Learning from Typed Feedback (Jev-DLC). Jev-DLC treats each correction as both evidence about a recurring action identity and a directed relation between the preferred and rejected choices. It preserves these signals in complementary bounded channels above a frozen scorer and combines their scores to shape the next decision. The method retains no raw interaction history, adds no scorer call, and performs no runtime language-model backward
标题：Asking Less, Adapting Better: Feedback-Efficient Test-Time Adaptation for Vision-and-Language Navigation
来源：arXiv
链接：https://arxiv.org/abs/2610.30643
摘要：Vision-and-Language Navigation (VLN) requires agents to follow instructions and navigate to target locations. However, current VLN agents often suffer severe performance degradation in unseen environments due to distribution shifts. Test-Time Adaptation (TTA) offers a promising paradigm to bridge this gap, yet existing methods that rely on unsupervised entropy minimization can propagate errors through the trajectory, causing cascading failures in navigation. Recent methods mitigate this by incorporating human feedback, but they apply trajectory-level supervision uniformly across all steps without localizing the actions is truly informative, leading to inefficient feedback utilization and mismatched optimization granularity. To address these limitations, we propose FETA, a Feedback-Efficien
标题：IQuest-Q1: Advancing Foundation Capabilities for Agentic CLI Systems
来源：arXiv
链接：https://arxiv.org/abs/2610.30949
摘要：We present IQuest-Q1, a sparse Mixture-of-Experts (MoE) agentic foundation model with 320B total parameters (15B activated). IQuest-Q1 is developed through a three-stage process: (1) stable and efficient pre-training, combining Stable SplitNorm, Muon, and muP with a two-stage curriculum whose second stage increases the coding-data mixture, preserving a wide stability margin between safe and optimal learning rates; (2) fine-grained mid-training, expanding the data mixture across domains and formats and progressively extending the context; and (3) multi-scaffold with multi-version post-training, combining SFT, domain-specialized RL, and MOPD across various scaffolds to integrate complementary training signals and strengthen agentic coding capabilities. Furthermore, model merging combines dif
[合并自 2026-10-24 iquest-q1-cli-320b-moe]
[IQuest-Q1：提升智能体 CLI 系统的基础能力（54 ▲）](https://huggingface.co/papers/2610.30949?utm_source=digest-papers&utm_medium=email&utm_campaign=2026-10-23)
[合并自 2026-10-24 iquest-q1-cli-moe]
标题：IQuest-Q1: Advancing Foundation Capabilities for Agentic CLI Systems
来源：arXiv
链接：https://arxiv.org/abs/2610.30949
摘要：We present IQuest-Q1, a sparse Mixture-of-Experts (MoE) agentic foundation model with 320B total parameters (15B activated). IQuest-Q1 is developed through a three-stage process: (1) stable and efficient pre-training, combining Stable SplitNorm, Muon, and muP with a two-stage curriculum whose second stage increases the coding-data mixture, preserving a wide stability margin between safe and optimal learning rates; (2) fine-grained mid-training, expanding the data mixture across domains and formats and progressively extending the context; and (3) multi-scaffold with multi-version post-training, combining SFT, domain-specialized RL, and MOPD across various scaffolds to integrate complementary training signals and strengthen agentic coding capabilities. Furthermore, model merging combines dif
[合并自 2026-10-25 iquest-q1-agentic-cli-moe]
标题：IQuest-Q1: Advancing Foundation Capabilities for Agentic CLI Systems
来源：arXiv
链接：https://arxiv.org/abs/2610.30949
摘要：We present IQuest-Q1, a sparse Mixture-of-Experts (MoE) agentic foundation model with 320B total parameters (15B activated). IQuest-Q1 is developed through a three-stage process: (1) stable and efficient pre-training, combining Stable SplitNorm, Muon, and muP with a two-stage curriculum whose second stage increases the coding-data mixture, preserving a wide stability margin between safe and optimal learning rates; (2) fine-grained mid-training, expanding the data mixture across domains and formats and progressively extending the context; and (3) multi-scaffold with multi-version post-training, combining SFT, domain-specialized RL, and MOPD across various scaffolds to integrate complementary training signals and strengthen agentic coding capabilities. Furthermore, model merging combines dif
标题：Joint or Per-Option? Compact Encoders for Closed-Set Decisions from 17 Million to 1 Billion Parameters
来源：arXiv
链接：https://arxiv.org/abs/2610.30970
摘要：Many language-model pipelines end in a closed-set decision, such as selecting a tool. Prompting a large language model for it is slow and order-sensitive. We ask how small an encoder for such decisions can be and whether it should read the options jointly or separately. In a three-seed study at 32 million parameters, per-option scoring beats joint scoring on a held-out sample of a 37-benchmark decision suite by 5.8 points of chance-corrected skill at matched optimizer steps and by 2.8 at approximately matched training compute; two external benchmarks agree, though not significantly on their unseen tasks. Crossing reader with initialisation shows that the advantage requires a reranker initialisation, whose query-candidate pair input matches per-option reading; from the raw encoder the two r
标题：RetiJev: Fast and Source-Grounded Ophthalmic Decision Models
来源：arXiv
链接：https://arxiv.org/abs/2610.30947
摘要：Ophthalmic clinical decision support requires fast, accurate, and privacy-preserving inference, yet autoregressive large language models (LLMs) incur substantial latency and computational overhead for decisions among predefined actions. We therefore investigate a decision-oriented alternative to generative reasoning. Specifically, we first introduce RetiJev-Data, a source-grounded ophthalmology corpus comprising 78,158 clinical-decision and diagnosis questions derived from articles in PubMed Central and a MeSH vocabulary covering 531 eye diseases, and 50,074 guideline-grading questions spanning 11 grading systems, along with a 24,662-question evaluation benchmark. Based on this, we develop RetiJev, a family of 0.8B-9B decision models that directly score candidate answers without autoregres
标题：Tracing Subliminal Learning Through Pretraining
来源：arXiv
链接：https://arxiv.org/abs/2610.30884
摘要：Large language models trained on data generated by other models can sometimes inherit traits unrelated to that data, a phenomenon called subliminal learning. We study both hard label distillation, where students train on numbers generated by a biased teacher, and on-policy soft label distillation, where students generate the data and the teacher's logits provide the learning signal. Using Pythia, PolyPythia, and DataDecide OLMo models, we find that transfer persists in both settings across differences in weight initialization, pretraining data order, pretraining corpus, model size, and architecture. Across pretraining checkpoints, the ability to transmit a preference and the ability to acquire one develop separately: some checkpoints are more effective teachers but weaker students, and vic
标题：The Value of System 1 Models in Medicine
来源：arXiv
链接：https://arxiv.org/abs/2610.31315
摘要：Many clinical decisions are a choice among a few fixed options, taken quickly and in large numbers. Such decisions also need an uncertainty estimate to know when to defer to a clinician. Generative language models answer slowly in free text and state their confidence poorly. In contrast, System 1 models answer zero-shot with a probability for each answer. However, their performance across clinical tasks is unknown. Four System 1 models, two proprietary and two with open weights, were evaluated zero-shot on 86 public tasks from 14 clinical domains in 20 languages, spanning examination questions and medical imaging. Tasks were assigned to six capability tiers before analysis. The primary outcome was the area under the receiver-operating-characteristic curve (AUROC). Secondary outcomes includ
标题：ReDence: A Hybrid Agentic and Decision-Model Framework for Rapid, Evidence-Grounded Detection of Modern Social Engineering Scams
来源：arXiv
链接：https://arxiv.org/abs/2610.30963
摘要：Detecting online social engineering (SE) scams often requires examining website workflows, verifying claims against external sources, and connecting evidence across multiple pages. Existing detectors provide limited support for these investigations, leaving analysts to manually gather and validate evidence, which inevitably delays detection and subsequent takedown of these attacks. To address this gap, we present ReDence, an open-source hybrid framework for adaptive, evidence-grounded detection across diverse SE scam categories. ReDence combines two locally hosted models with complementary responsibilities: a fine-tuned multimodal agentic Investigator (Qwen3.5-4B) that performs investigative planning, evidence interpretation, and reporting, and a specialized Decider (Tev1) that handles bin
标题：BeNice! Closing Looping's Inference-Memory Gap via KV-Sharing
来源：arXiv
链接：https://arxiv.org/abs/2610.31633
摘要：Looped architectures scale quality while keeping parameter count fixed by repeatedly passing activations through the same block of layers. During inference, each recurrence generates independent KV-caches, causing memory to grow linearly. To address this challenge, prior work has proposed KV-sharing (i.e., sharing KV-caches across depth), finding that, in some cases, the KV-cache of the final recurrence can be shared with the earlier recurrences without quality loss, a property we denote as KV-shareability. However, the mechanisms underlying KV-shareability remain poorly understood, with even identically trained looped architectures not consistently exhibiting it. In this paper, we study how KV-shareability can be enabled in looped models. We propose a two-stage method: in the first stage,
标题：Compiling Autoregressive LLM Control into Policy-Exact Parallel Inference for Hierarchical RAG
来源：arXiv
链接：https://arxiv.org/abs/2610.30664
摘要：Large language models increasingly act as controllers in structured decision pipelines, yet their decisions are still executed by autoregressive text decoding. Hierarchical RAG is one such setting: at each frontier of a document tree, an LLM commits to $b$ of $k$ sections under an input-dependent action grammar, executed as either failure-prone free-form JSON or reliable but strictly serial grammar-constrained generation. We show that this serialization is unnecessary for bounded-horizon control: the reachable future forms a shallow action-prefix tree, which we compile --- one root prefill plus one tree-causal probe evaluate every prefix-conditioned action distribution with serial decoding's tokens, positions, and causal ancestry. A prefix-state lemma makes compiled control policy-exact in
标题：Drift Says When, Budget Says How Much: Rate-Controlled KV Re-selection for Long-Form Multimodal Generation
来源：arXiv
链接：https://arxiv.org/abs/2610.31433
摘要：KV cache compression for multimodal LLMs usually selects the visual tokens to keep once, at prefill, and reuses that selection for the whole generation. We show that this one-shot selection is severely limited on long-form generation: on detailed image description, even an oracle that selects with the attention of the uncompressed answer reaches only 0.44-0.52 ROUGE-L agreement with the uncompressed output, because the evidence the model reads moves as the description unfolds. Re-selecting whenever attention drifts out of the kept set is the natural remedy, but we find that a fixed drift threshold fails systematically: drift magnitudes are not comparable across samples, so low-drift samples are almost never refreshed although they benefit the most. We propose Rate-Controlled Refresh (RCR),
标题：RIFT: KV Cache Compression for Draft Models with Complementary Compressibility Training
来源：arXiv
链接：https://arxiv.org/abs/2610.31098
摘要：Speculative decoding is a popular decode acceleration method. As such, it is particularly relevant to long-output generation, which requires an increasingly large KV cache. KV cache compression offers a way to constrain costs in long-context drafting. However, compression methods that focus on prefill information can lose relevance as the output length grows. We term this phenomenon decode staleness and show that it degrades draft model accuracy under selection-based KV cache compression. We introduce RIFT, a family of selection-based KV cache compression strategies that balance prefill information and recent generated tokens. We show that refreshing the KV cache selection during token generation significantly improves the drafter's quality at the cost of additional decode latency. To alle
标题：HERD: Harness-Evolving Recursive Distillation for Continual Learning
来源：arXiv
链接：https://arxiv.org/abs/2610.31401
摘要：Continual learning requires a model to keep acquiring new skills and knowledge without compromising its existing capabilities. Parametric methods suffer from catastrophic forgetting and loss of plasticity. Non-parametric methods, such as prompts, skills, and harnesses, leverage in-context learning but are limited by context length and lossy plain-text compression. To effectively combine their complementary strengths, we propose HERD (Harness-Evolving Recursive Distillation), a hybrid continual learning framework that learns efficiently in the harness and internalizes experience into cumulative capabilities via distillation. On the harness side, HERD continually refines its skills to capture reusable patterns, while retaining high-residual examples that the skills fail to capture in a bound
标题：RIDAR: Safeguarding Research Integrity in Autonomous Research Agents via Automated Detection and Runtime Intervention
来源：arXiv
链接：https://arxiv.org/abs/2610.31372
摘要：Autonomous research agents are increasingly used to automate large portions of scientific research, yet they are prone to research integrity violations (RIVs): behaviors that compromise the validity, reproducibility, or faithfulness of the research process. We present RIDAR, a framework for automated RIV detection and runtime intervention. First, we introduce RIVT, a comprehensive taxonomy of 14 RIV types spanning the full research lifecycle. Building on RIVT, we present RIDAR-post-hoc, a framework to detect RIVs within agent trajectories. RIDAR-post-hoc achieves 85.2% F1 performance on RIV-Gold, a benchmark for RIV detection with 60 expert-annotated trajectories. Applying RIDAR-post-hoc to RIV-Wild, a dataset of 300 agent trajectories across 50 distinct research tasks, we detect 3,964 RIV
标题：Subgoal-aware Memory Graph for GUI Agent
来源：arXiv
链接：https://arxiv.org/abs/2610.31307
摘要：GUI agents aim to automate complex user tasks by interacting with mobile interfaces, but ef- fective decision-making in long-horizon set- tings requires reusing prior experience rather than relying solely on the environment ob- servations. Existing memory-based methods typically store experience either at the single- action level or as entire trajectories. Single- step actions are often too fine-grained to carry meaningful semantics, since the success of an individual action does not necessarily imply progress towards the overall goal. In contrast, whole trajectories are often too coarse-grained and difficult to reuse across different tasks. We propose a subgoal-aware memory graph that distils past trajectories into structured, reusable evidence at the level of subgoals. We define a subgoa
标题：The Value of Memory in Language Agents
来源：arXiv
链接：https://arxiv.org/abs/2610.31303
摘要：Memory enables language agents to carry useful distinctions from past experience into future decisions. Its value depends on which distinctions survive, whether a bounded agent can use them, and what can be recovered or revised as demands evolve. We develop a decision-theoretic framework that follows this value from raw history through persistent state, operational use, and costly recovery and revision. Original and retained value are expressed in units of avoided task loss; risk-cost frontiers describe attainable recovery and preservation across uses. Unknown demand yields value profiles, while interactive feedback changes the distribution of future work. A cost-aware simulation condition and five minimal counterexamples expose the roles of information timing, retrieval access, component
标题：Governance-State Corruption in Long-Term Memory for LLM Agents: A Cross-Model Stress Test
来源：arXiv
链接：https://arxiv.org/abs/2610.30698
摘要：An LLM agent can retrieve a memory correctly yet use it incorrectly after its validity state changes. We study governance-state corruption: faults that alter which retained records enter model-visible context while stored content and tasks remain fixed. Across six providers, 731 benchmark-grounded bases yield 2,080 matched fault pairs. In offline control-plane checks, corruption changes the operative projection in 96.25% of pairs, while oracle-backed enforcement restores it in 100%. For downstream evaluation, a deterministically selected, non-probability LongMemEval/Memora panel resolves 298 of 300 matched triplets (238 distinct bases). Clean-minus-corrupted contrasts are 40.54 percentage points in LongMemEval binary correctness (95% base-cluster interval [30.32, 50.71]) and 30.60 points i
标题：ColaBench: Evaluating Memory-Guided Knowledge Acquisition in Personal Agents
来源：arXiv
链接：https://arxiv.org/abs/2610.30989
摘要：Personal agents need to draw on long-term interaction history to retrieve and apply external knowledge suited to a user's current request. Even an answer grounded in relevant documents can be unsuitable when it overlooks the user's circumstances. This creates two connected challenges: using historical conditions to identify the knowledge needed for a new request, and applying that knowledge to the user's situation. Existing benchmarks advance memory recall, personalized retrieval, and memory-guided action, while the joint evaluation of these two challenges remains underexplored. We introduce ColaBench, a synthetic benchmark in health and daily living, with 349 tasks across 10 Profiles and a knowledge base of 13,033 documents. Its Evidence Graph connects personal conditions to their disclos
标题：Memory-Aware Reward Modeling: Learning Human Preferences for LLM Memory Use
来源：arXiv
链接：https://arxiv.org/abs/2610.31272
摘要：Memory enables language model assistants to maintain continuity across sessions and generate personalized responses, but knowing more about a user does not necessarily produce responses they prefer. Existing evaluations assess memory capabilities without directly establishing whether memory use benefits the user. To study this gap, we collect approximately 1,000 pairwise judgments from 87 participants using their own ChatGPT queries and personal memories. We find that memory-conditioned responses are not consistently preferred over responses generated without memory and in some cases, users strongly prefer the latter. Automatic evaluators also show limited agreement with users and favor memory-conditioned responses substantially more often than users do. Motivated by these gaps, we introdu
标题：Delta N-gram: Forward-Only Continual Writing into a Language Model's Native N-gram Memory
来源：arXiv
链接：https://arxiv.org/abs/2610.31125
摘要：Our previous analysis found that a model's native hashed N-gram table is where continual updates disturb the base least, but its writes needed backpropagation and changed the pretrained table. We present Delta N-gram, a forward-only write path into that table. The pretrained table T^0 is never modified. Evidence is written into a sparse patch Δon the rows it addresses: a 4.1M-parameter writer emits a target and a gate per (position, hash head) event, and a block-averaged delta rule keeps each row a convex combination of its old value and the gated targets. The writer is trained offline on raw text with the pretraining objective; at use time writing takes forward passes only, and clearing Δrestores the original model. A 0.6B prototype trained to answer questions keeps 16/16 facts after a se
标题：RIPPLE: Recovery Internalization through Process-Level Policy Learning and Effect-Aware Credit for Long-Horizon Agentic Coding
来源：arXiv
链接：https://arxiv.org/abs/2610.31276
摘要：Long-horizon coding agents can observe a failure and generate a plausible critique yet continue along the same incorrect trajectory. We call this the Critique-to-Action Handoff Gap. We introduce RIPPLE, a recovery-oriented training framework that internalizes process-level guidance into a single self-recovering policy. A prefix-only teacher provides evidence-grounded recovery supervision without access to future actions or terminal outcomes. After lightweight distillation, reinforcement learning assigns stage-specific credit to reflection triggering, diagnosis, and corrective action. Local validity signals gate trajectory advantages so that intermediate recovery decisions do not indiscriminately inherit terminal outcomes. Across four repository-level software-engineering benchmarks, RIPPLE
标题：Directing to Diversify: Multi-Directional Search for Automatic Heuristic Design
来源：arXiv
链接：https://arxiv.org/abs/2610.31048
摘要：Large language models (LLMs) have demonstrated strong potential to automate heuristic design for NP-hard combinatorial optimization (CO) tasks without human intervention. Recent methods predominantly rely on evolutionary computation (EC) frameworks, where heuristic diversification is mainly induced through EC-like operators or downstream selection, while the direction of heuristic transformation remains largely unharnessed. In this paper, we propose a multi-directional Perturbation-Rewrite Search (PRS) framework for automatic heuristic design that factorizes heuristic transformation through directional search primitives at the generation level. Specifically, PRS operates on heuristic ideas through three complementary directions: expansion toward richer and more expressive designs, compress
标题：CANNBot-CAKE: Collaborative Agentic Kernel Evolution for Ascend NPU
来源：arXiv
链接：https://arxiv.org/abs/2610.31034
摘要：Efficient custom operators are essential for modern deep learning workloads, yet developing high-performance kernels for Ascend NPUs remains slow, expert-intensive, and difficult to automate. Although AI agents now approach expert performance on GPU kernel benchmarks, current models remain less capable of generating AscendC kernels for NPUs, potentially reflecting the relative scarcity of relevant training data. We thus present CAKE, a CANN-native system for deployable custom operators that combines adaptive multi-DSL routing, separate Generation and Evolution modes, and reviewed knowledge reuse through CAKE Book and a Self-Evolving Agent. On 53 CANN-Bench v1.1 operators on Ascend 950PR, CAKE's aggregated results achieve $100\%$ case pass rate on the standard set (1,060/1,060) and $98.28\%
标题：Learning What to Ask for Tomorrow: Active Preference Acquisition for Long-Term Personalized Web Agents
来源：arXiv
链接：https://arxiv.org/abs/2610.31107
摘要：Web agents increasingly serve the same user across repeated online tasks such as booking travel, shopping, and scheduling, so their long-term usefulness depends on how well they know the user's preferences. Passive memory and personalization methods cannot obtain preferences the user never reveals, and clarification methods that value a question by its benefit to the current task overlook the benefit of reusing the answer in later tasks. We study Active Preference Acquisition (APA): which missing preferences are worth the cost of asking? We define the Prospective Personalization Value (PPV) of a missing preference as its expected cumulative utility gain over future tasks minus the effort and disclosure-risk cost of asking, and select questions under a per-interaction budget. We characteriz
★2Today AI 累计融资超 2 亿美元，布局主动式个人智能体未知 · 产业动态分析 · 原文Today AI 宣布累计完成超 2 亿美元融资，并以长程记忆、主动触达和跨应用任务执行能力打造个性化个人智能体。
标题：ReflectPO: Reflection-Guided Policy Optimization for Long-Horizon LLM Agents
来源：arXiv
链接：https://arxiv.org/abs/2610.30933
摘要：With external tools, large language models (LLMs) can tackle complex tasks through sustained environmental interaction. LLM-based agents are therefore increasingly used to tackle complex long-horizon tasks in real-world environments. However, these tasks involve sequential, interdependent decisions and changes in environmental states. Early execution errors affect subsequent interaction histories and environmental states, and their effects accumulate along the decision chain. To address this challenge, we propose Reflection-Guided Policy Optimization (ReflectPO), a reinforcement learning framework for autonomous error recovery in long-horizon tasks. ReflectPO comprises two core components. The first is Reflection-Guided Sampling. We use branch differences and repeated visits in interaction
标题：Training Small Language Models for Transferable Tool Selection
来源：arXiv
链接：https://arxiv.org/abs/2610.31483
摘要：As LLM agents operate over corpora of more than ten thousand tools, selecting the right tool for a complex multi-step user instruction becomes a central bottleneck. Tool descriptions cannot all fit in context, and relevance-based retrievers only rank tools without knowing which are actually needed. Recent work instead leverages proprietary commercial LLMs like Claude or ChatGPT to drive a retriever, decomposing the instruction and selecting over retrieved candidates. We show that a small, open model can be trained via reinforcement learning to perform the same functionality. On ToolBench, a Qwen3-8B model post-trained with our recipe outperforms Llama-4-Maverick (402B). In addition, the trained model transfers directly to different retrievers and generalizes to new tool corpora in differen
标题：Separate Field Scoring Protects Unused Tools When Agents Pool Tool-Use Experience
来源：arXiv
链接：https://arxiv.org/abs/2610.30867
摘要：Agents that call tools from a shared catalog can pool what they learn about which tool serves which request. We measure what pooled usage examples do to tool retrieval, reporting separately the held-out tools, for which no agent has usage, and all other tools. Adding usage examples to the index as extra entries for used tools raises Hit@5 (whether a gold tool is in the top five) for those tools and lowers it for held-out tools, by a large margin on two of three catalogs and with both embedders we tested. Giving used tools the same number of equally long entries filled with other tools' text lowers held-out Hit@5 about as much, but gives little or none of the gain. We call our method of scoring usage as its own field SYNAPSE. On StableToolBench, the held-out gold tool appears in its top fiv
标题：Adaptive Skill Granularity for Coding Agents: Runtime Routing and Macro-to-Micro Recovery
来源：arXiv
链接：https://arxiv.org/abs/2610.31461
摘要：Reusable skills allow large language model agents to package task-specific procedures, tool affordances, and execution guidance into modular units. Prior work has largely studied how agents retrieve, compose, or synthesize skills before execution. In coding agents, however, execution poses an additional runtime control problem: the same task may be solved through primitive tool calls, macro skills, or hybrid execution that starts with a macro and falls back to micro-level recovery. We introduce Adaptive Skill Granularity, a framework that treats skill granularity as an action-abstraction policy conditioned on task structure, execution state, and macro-skill failures. The system combines a feature-based granularity controller and a finite-state recovery executor. We evaluate the framework w
标题：GIFT: Gradient Informed Filtering for Distillation Augmented Reasoning RL
来源：arXiv
链接：https://arxiv.org/abs/2610.30796
摘要：Training large language models for complex reasoning tasks relies on two complementary signals: sparse, outcome-grounded reinforcement learning (RL) for correctness, and dense, token-level distillation for guidance. However, complementarity in supervision does not guarantee compatibility in optimization; imitation-focused distillation can conflict with the outcome-driven RL objective, and existing hybrids lack a principled resolution. We propose \textbf{GIFT} (\textbf{G}radient-\textbf{I}nformed \textbf{F}il\textbf{T}ering for Distillation-Augmented Reasoning RL), a novel method that directly addresses the allocation problem of which teacher updates to apply. GIFT performs a token-level geometric analysis, decomposing the distillation gradient into a component \textbf{parallel} to the RL g
标题：One Teacher, Many Views: Reducing Reference Sensitivity in On-Policy Self-Distillation
来源：arXiv
链接：https://arxiv.org/abs/2610.30880
摘要：On-policy self-distillation (OPSD) improves reasoning by training a model on its own trajectories with dense token-level supervision from a privileged copy conditioned on a reference solution. Existing methods typically use a single teacher distribution conditioned on the complete reference. We identify a limitation of this formulation: a reference is only one realization of the reasoning process. Although it provides evidence about the answer, it also encodes an ordering, phrasing, and solution path. Full-reference conditioning can therefore produce correct targets that depend excessively on information unavailable to the student at inference. We show that replacing one valid reference with another, while holding the problem and student state fixed, can substantially alter the teacher dis
标题：Who Is Sycophantic to Whom? Confident LLMs, Human Deference, and the Collapse of Trust Calibration
来源：arXiv
链接：https://arxiv.org/abs/2610.31579
摘要：Research on sycophancy asks whether large language models (LLMs) bend towards what their users want to hear. We ask the reverse question: when an LLM is confident, consistent, and wrong, do users bend towards it? In a within-subjects experiment, 32 participants learned up to five science topics by chatting with five assistants built on the same model (GPT-5-mini). Unknown to participants, each assistant's briefing had 0, 25, 50, 75, or 100% of its facts replaced by plausible falsehoods, and every assistant spoke with full confidence. A surprise quiz after each topic offered the correct fact, the planted falsehood, and two distractors, with a confidence rating for every answer. Across 104 quizzes and 1,040 answers, scores fell from 7.4 to 2.9 out of 10, and the planted option was chosen on
标题：Fine-Grained Analysis of Catastrophic Forgetting in Continual Learning via Algorithmic Stability
来源：arXiv
链接：https://arxiv.org/abs/2610.31551
摘要：Understanding catastrophic forgetting, in which learning new tasks degrades performance on previously learned tasks, remains a fundamental challenge in continual learning (CL). Existing theoretical analyses often rely on specific model or task structures and offer limited insight into how iterative optimization affects forgetting. To address this gap, we develop an algorithmic stability framework for expected population forgetting under SGD without prescribing a particular model architecture. Specifically, we decompose forgetting into intra-task and cross-task generalization gaps and empirical forgetting. To bound the generalization gaps, we track how a sample perturbation affects updates within a task and propagates through subsequent tasks. We consider constant, intra-task decaying, and
标题：Trace2World: Reconstructing App Worlds from Real Interaction Traces for Scalable GUI Agent Training
来源：arXiv
链接：https://arxiv.org/abs/2610.30792
摘要：Scaling GUI reinforcement learning requires environments that preserve real-app behavior. Yet coupling behavior inference with coding can turn partial traces into runnable mocks with incorrect state transitions. We introduce \textbf{Trace2World}, which separates global behavior reconstruction from implementation. It reconciles traces into a shared, source-linked app model, then compiles an executable environment. Independent replay of source workflows localizes mismatches for bounded implementation repair. Validated worlds then support learner-adaptive task and environment augmentation within source-supported behavior. From 1,570 traces, Trace2World constructs 50 environments. On 200 workflows across ten apps, 91.1% are reproduced end to end, versus 59.8% for the strongest matched task-loc
标题：Network topology shapes consensus and vulnerability in LLM populations
来源：arXiv
链接：https://arxiv.org/abs/2610.31165
摘要：As large language models (LLMs) are deployed in interacting populations, their collective behaviour depends not only on the models but also on who interacts with whom. Studies of convention formation in LLM populations have mainly tested well-mixed interactions. We place policies extracted from four LLMs on fully connected, random regular, and heterogeneous networks, with $10^3$ to $10^5$ agents playing the naming game. Topology affects how quickly consensus emerges, sometimes whether it emerges at all, and, in rare cases, which convention wins. Sparsest networks slow consensus up to twelvefold. Yet, by preserving correlations between persistent neighbours, they restore consensus in some cases that fail to converge under well-mixed interactions. Networks also govern vulnerability to advers
SpatialScout 是面向视觉语言模型智能体的主动式物体级三维建图基准，包含80个多样化室内场景和3140个目标物体。智能体根据物体清单、RGB观测、相机标定及位姿控制相机探索场景，并在统一坐标系中估计物体中心。任务综合考察视点选择、跨视图关联、几何估计和地图修订，并以兼顾定位精度与目标覆盖率的物体尺度重叠分数进行评价。实验中表现最强的 Claude Opus 5.5 平均重叠分数仅为41.71%，说明当前前沿VLM距离构建完整、精确的度量地图仍有显著差距。
