Papers

Alignment applied after pretraining is shallow in a measurable way: a single direction in a model's residual stream can be edited out, and the model stops refusing harmful requests. That fact says how easily refusal can be removed, not what the refusal decision was reading in the first place. We ask what it reads, and we separate that from what the model comprehends. Across four open-weight models spanning three families, moral comprehension is native to pretraining: a low-rank moral subspace crystallizes during pretraining, and alignment rotates it once without rebuilding it. The refusal gate, in contrast, is a fresh post-training construction with only a weak pretraining precursor, written into a narrow control-token channel where the refusal decision is orthogonal to the moral-judgment decision. The central result is causal and comes from one model, OLMo-3. A nested interchange rank sweep patches successively larger slices of the moral subspace between matched requests and reads how much of refusal's response transfers: as the basis widens, moral judgment keeps reading more of it, while refusal levels off at the level of a single harm direction, and about three-quarters of refusal's causal input lies outside the moral subspace altogether. Refusal reads the harm percept, not the moral content that judgment reads on the same patches. The picture is not uniform across families. Llama reads broad moral content; Qwen reads beyond the single harm cue but is unresolved at our sample size; GPT-OSS reads harm, and its refusals can be argued in either direction by its own reasoning trace. Where refusal reads only a low-rank slice and routes around the bulk of what the model knows, a rank-one edit removes it. Whether widening what refusal reads would also deepen the behavior is the open question this raises.

Tool-augmented language models are evaluated on whether they reach the right answer, not on whether they report honestly when a tool fails to supply one. We isolate this post-failure decision with a benchmark of 1,024 items spanning 16 internal-system domains and eight tool-failure types, in which a tool call is enforced and the returned payload is guaranteed to be unusable. Under a deployment-style system prompt, 14.10% of responses are dishonest: the model either asserts a value the payload cannot support or declines while citing a fabricated policy or capability limit. The rate is governed almost entirely by whether the failure is signalled. When the tool returns status:error, dishonesty is absent (0.0%); when it returns status:ok with a redacted, corrupted, stale, malformed, empty or truncated value, dishonesty reaches 45.3%. The behaviour is not an artefact of our prompts: it appears under a neutral prompt (10.17%) and under the shipped prompt of every production agent framework we evaluate, reaching 24.67% under CrewAI's, and none of the nine frameworks we audit specifies what the model should do when a tool fails. Comparing prompt-level defences, we find that the operative variable is not deference to tool output but the absence of a named failure state. Appending a single sentence that requires the model to emit retrieval_status: OK or FAILED before answering reduces dishonesty from 14.10% to 0.87%, with one item of 688 worsening against 92 improving, and transfers unchanged into three foreign agent scaffolds. The emitted flag is faithful in 99.7-99.9% of declarations, giving a runtime detector that needs only a regular expression.

Causal claims about large language model (LLM) internals rest on measurements. Those might include a projection, a cosine, an ablation delta, or an interchange patch among others. These measurements fail in specific, diagnosable ways that return a plausible number instead of an error, so a broken instrument can easily read as a finding. A covariance-matched null can saturate until every direction looks typical, a per-head attribution can overshoot the true residual write threefold on reordered-normalization architectures, an interchange patch can go sign-chaotic because its outcome is pinned at a ceiling, or a read-from verdict can be an artifact of measuring past the layer where the model already decided. This note documents six such failures from a causal interpretability program on refusal and moral representation, spanning several papers and a four-model open-weight panel; each mode is established on one or two of the four. For each we give the tell that catches it and a protocol keyed to a detectable trigger (reordered normalization, massive activations, a low-dimensional decision channel), so we and readers can check whether a given setup is exposed. The discipline reduces to four moves: calibrate against a positive-control ladder, certify with an orthogonal cell, compute power before spending compute, and state every read-from verdict at a depth referenced to the model's commitment. The evidence is four architectures across three families within a single program; external replication across programs is future work.

Pure speech language models often lag behind text and speech-text language models in generating coherent content, but this gap is difficult to quantify because speech and text systems are typically evaluated with different metrics and trained on different data. We study the speech-text modality gap in a family of spoken language models, based on flow matching for continuous acoustic feature generation. We construct a unified generation-based evaluation suite that compares speech-only, text-only, and speech-text language models trained on matched data distributions and evaluated in matched generation settings. We evaluate generated continuations along multiple dimensions: semantic coherence, measured by transcribing generated speech and scoring it with a reference language model; local phonetic structure, measured by phone n-gram distributional statistics; speaker consistency and acoustic quality; and emotion-based distributional metrics. Across datasets, we find that joint speech-text modeling substantially improves semantic coherence. However, the improvement is not uniform across metrics: phone-level metrics change only modestly, speaker similarity and predicted quality are lower for speech-text continuations, while emotion-based distributional metrics improve. Compared with larger-scale speech-only models, our speech-text model closes much of the scaling gap in transcript-based semantic coherence, suggesting that text provides an efficient semantic training signal for spoken language modeling.

Large language models are increasingly used in high-stakes domains such as law, where systems must ground their reasoning in retrieved evidence and abstain when that evidence is insufficient. However, existing reward models are largely optimised for general preferences rather than contextual grounding, limiting their ability to evaluate these behaviours in retrieval-augmented generation (RAG) settings. We introduce a framework for transforming existing legal QA datasets into contextual preference data and use it to construct LegalRewardBench (LRB), a benchmark for evaluating grounded legal generation under noisy and insufficient retrieval conditions. Across general and legal contextual evaluation, we find that contextual DPO improves grounded evaluation, but performance is sensitive to preference-data construction. Length-balanced augmentation substantially improves grounded legal evaluation, with the strongest configuration combining length-balanced legal and general contextual preference data and improving performance by up to $\mathbf{+25.6}$pp over baseline. We further find evidence of cross-jurisdiction transfer: models contextually refined primarily on Victorian criminal-law data improve grounded evaluation on external US legal benchmarks, including a $\mathbf{+16.2}$pp improvement on Housing Statute QA. Together, these results provide a reproducible foundation for constructing and evaluating grounded legal reward models in retrieval-augmented settings.

Large language models are increasingly used to scale codebook-based annotation in scientific research, but existing workflows provide limited support for translating domain experts' codebooks into reliable, revisable, and auditable prompts. Prompts are often treated as fixed instructions and hidden from annotators, making it difficult for non-technical domain experts to diagnose and correct model behavior when outputs violate codebook guidelines. In this paper, we present CALICO, a human-centered, codebook-aligned annotation workflow that treats prompts as editable, versioned, and optimizable artifacts. CALICO integrates codebook parsing, prompt generation, result inspection, prompt versioning, natural language human feedback, and label-supervised prompt optimization through existing optimizers such as GEPA, MIPROv2, and OPRO, together with our reflection-based optimizer, ReflectAgent. Empirically, we evaluate CALICO on domain-specific AI-companion chatbot conversation codebooks. Across evaluated dimensions, CALICO improves mean held-out performance by +13.0 and +7.4 absolute points for two coders, respectively. A coder-specificity analysis further suggests that optimized prompts capture coder-specific interpretations rather than only generic codebook clarification. CALICO runs as a web application that takes users from raw codebook materials to inspectable, exportable labels; the website, codebase, and live demo are released at https://calico-annotation.github.io/ under the Apache 2.0 License.

We scale our conventional sub-150M pretraining recipe from 53.5M to 109.7M parameters, holding the method fixed (Qwen3-style decoder with grouped-query attention, RoPE, SwiGLU, RMSNorm, QK-Norm, and a z-loss; FineWeb-Edu data) and changing only the geometry to a deep-and-thin 23-layer x 576-hidden design. The larger model improves across the board – BLiMP 78.1 -> 81.3, ARC-Easy 51.4 -> 52.5, WikiText-2 byte-perplexity 2.04 -> 1.95 – and its 81.3% BLiMP essentially matches GPT-X2-125M (81.28) at about 12% fewer parameters. Notably the 110M model achieves this on fewer training tokens (about 8B vs 12B), so the gain is attributable to capacity and depth, not more data. Both models are deliberately conventional; this report is a clean scaling control and the baseline rung (R0) of an ablation study of what further improves models in this regime. An ablation ladder follows: value residuals (R1) and the Muon optimizer (R2) lift ARC-Easy by a cumulative +3.6 (52.5 -> 56.1) at a near-flat BLiMP and are kept; a diverse data blend (R3) and two logit-distillation settings (R4a/R4b) are not kept – honest negatives. R3 pins ARC-Easy to FineWeb-Edu's educational filtering rather than raw diversity; distillation from a 1.7B teacher can reach the class-leading ARC-Easy (56.99, matching GPT-X2-125M) but only at a perplexity cost that dialing KD down then erases – so R2 remains the best kept stack.

As autonomous vehicles and Extended Reality (XR) headsets enable novel in-car interactions, seamlessly querying physical landmarks, known as Outside-the-Vehicle Referencing (OVR), remains challenging due to ego-motion and referential ambiguity. We present a robust, multimodal OVR framework fusing user gaze and natural language to identify Points of Interest (POIs). To address the scarcity of dynamic vehicular data, we developed a VR-based pipeline synchronizing 360-degree transit videos with vehicle GNSS telemetry. Through a user study (N=46) mapping passenger head orientation into a 3D geospatial Digital Twin, we captured authentic gaze-speech behaviors. We subsequently trained a lightweight Transformer network, leveraging LLMs to dynamically align continuous spatial gaze vectors with discrete verbal context. Experimental results demonstrate high accuracy and low computational overhead, achieving an 83.33% Top-1 accuracy (87.72% Top-2) and an average inference time of 24.3 milliseconds. This real-time paradigm effectively resolves referential ambiguity, enabling context-aware spatial retrieval for passengers within the vehicle.

Data-to-text driver coaching is often presented as a generic pipeline from telematics events to advice. This paper argues that its content requires localisation because usefulness and credibility depend on drivers' knowledge, prevalent risks, regulation, infrastructure, and available data. Two independently developed systems in the United Kingdom and Nigeria are compared by tracing requirements through content selection, generation, and field evaluation. The UK system prioritises post-trip reflection, explanations tied to road and place context, and tone-sensitive wording. The Nigerian system combines legally grounded, once-daily Tips based on detected events with weekly persuasive Reports; it foregrounds safety education and alcohol-related risk in response to reported gaps in formal training and traffic-rule knowledge, as well as local road-safety priorities. Reliable speed-limit metadata supported speeding feedback in the UK, whereas its scarcity led the Nigerian evaluation to exclude speeding from its outcome metric. Both interventions were associated with reduced distance-normalised unsafe-event rates in their own field studies, although their designs and metrics preclude an effect-size comparison. The analysis yields a requirements-to-content design process for localising safety-critical NLG without treating a high-income deployment as the default.

Large language models (LLMs) have changed the way people engage with stories. Drawing on public chatbot logs, we can see that when users generate stories, they iteratively edit their prompts to explore narrative possibilities, adjusting characters, redirecting plots, and swapping fictional universes. As aggregated data, these prompts represent rich traces of creative preference at scale. Yet story generation evaluation benchmarks rely on static, one-shot prompts that cannot capture this exploratory behavior. In this work, we study how users revise consecutive story prompts in the wild. Using a dataset of naturally occurring user-chatbot conversations, we construct WildStories, a sample of 275,635 story generation prompts (labeled with story format, prompt components, and explicitness), and WildEdits, a collection of 24,291 edit trees that model how users iteratively edit base story prompts and explore branching story possibilities. From these trees we develop a framework of edit types crossing four directions (adding, removing, changing, and extending) with fourteen targets (e.g., plot, character, genre). We then use our datasets and this framework to analyze user behavior in navigating narrative space via LLMs. Finally, we show how automated permutations based on the framework can be used for story generation benchmarking. Content Warning: This paper works with "wild" chatbot logs, which often include toxic and sexually explicit themes.

Turn-taking is a fundamental component of spoken interaction, and while humans naturally rely on both verbal and non-verbal signals, dialogue systems usually depend on audio cues alone. This paper investigates whether visual features from face-to-face conversations can enhance turn-taking prediction beyond what is achievable from audio-only. We extend the Voice Activity Projection (VAP) model, a self-supervised transformer-based model for predicting future voice activity, by incorporating visual features extracted from the large-scale Meta Seamless Interaction dataset of dyadic face-to-face conversations. The visual features include gaze direction, head movement, body and hand pose, and facial action units (FAU). For incorporating the visual features, we explore concatenation, cross-attention fusion, delta features, and trainable gating mechanisms. Results show that visual information improves performance over the audio-only baseline, with FAU being significantly more informative than other feature groups. Body and gaze features nevertheless contribute complementary information, as the model combining all features performs best. Furthermore, results indicate that performance on specific tasks varies depending on whether training and test data come from improvised (acted) or naturalistic (non-acted) conversations.

Aligning multi-turn dialogue agents is usually framed as matching turn-level human preferences, yet direct optimization of long-term outcomes is often ineffective and prone to reward hacking. We formulate long-horizon dialogue optimization as a multi-objective reinforcement learning problem and train a multi-head value model that predicts a vector of observed user behaviors across multiple look-ahead horizons. Our findings demonstrate that a scalarized composite of dense auxiliary behavioral signals enables effective credit assignment and optimization of sparse outcomes. However, optimizing unconstrained single-objective proxies might induce policy degradations that are harmful when the agent is exposed to real users. To identify these failure modes prior to deployment, we establish a safety framework combining counterfactual user simulation with a validated dialogue-level outcome model to evaluate preference weightings and policy optimization methods. Finally, we demonstrate that distilling multi-objective value preferences into the policy via reference-anchored preference optimization matches on-policy online RL at a small fraction of its compute budget. Live A/B testing confirms that our distilled policy significantly improves long-term user retention, while simultaneously enhancing the positive behaviors and therapeutic-process markers.

Multi-turn interactions with LLMs are becoming increasingly common in information-seeking scenarios. However, user queries are often ambiguous and context-dependent, making them ill-suited for direct use as retriever queries. Conversational query reformulation (CQR) addresses this issue by rewriting the current utterance into a stand-alone query grounded in the dialogue history. Recent LLM-based CQR approaches achieve strong performance; however, their repeated LLM invocations and misalignment with downstream retrievers remain challenges. In this work, we begin from the observation that retrievers are highly sensitive to content ordering: simply reordering the same content can lead to changes in retrieval coverage and performance. Based on this, we propose a novel training-free method that generates a very large number of queries with minimal LLM usage by compositionally combining a small set of atomic components. We further apply LLM reasoning to construct a high-quality document set that balances precision and recall while capturing the user's core intent. Our framework generalizes across both open- and closed-source LLMs as well as dense and sparse retrievers. It achieves strong performance on four widely used conversational benchmarks, with up to 22.5% relative MRR improvement over the previous state-of-the-art baseline with far fewer LLM calls.

On-policy distillation (OPD) has become a standard approach for transferring capabilities from large teachers to compact students. Its cost, however, is dominated by autoregressive student rollouts and scales poorly in multi-turn agentic settings. Existing acceleration methods truncate or relocate the supervision signal according to fixed, offline budgets, despite substantial variation in teacher-signal reliability both within and across trajectories. Our empirical analysis on $\tau^2$-bench reveals a clear structure in this variation: informative supervision is concentrated in the prefix of each turn, and, most importantly for multi-turn agentic training, the cross-turn loss of teacher endorsement is temporally locked to the student's first erroneous action rather than accumulating gradually over turns. Building on these findings, we propose STRIDE (Stop-and-Restart on-policy Distillation acceleration), which combines two complementary techniques: adaptive early stopping, which terminates a rollout once the cumulative teacher log-probability falls below an out-of-distribution threshold, and a prefix buffer, which caches high-quality prefixes and restarts generation at the weakest correct turn. Together, these mechanisms induce a data-driven curriculum that progressively extends coverage to later turns. On $\tau^2$-bench retail, our method matches full-trajectory OPD and exceeds the 30B teacher at a $3.73\times$ speedup, surpasses the baseline itself at $2.34\times$, and retains a $4.51\times$ speedup under cross-domain multi-teacher training. As a supplementary generalization test beyond the agentic setting, STRIDE outperforms full OPD on AIME 2025 at a $5.10\times$ speedup and on AIME 2024 at a $3.08\times$ speedup; averaged across the two evaluations, both fixed-budget truncation baselines remain below full OPD.

This study highlights the role of domain-specific pretraining profile (DSPP) in Transformer performance for modeling digital pragmatics in Arabic-English code-switched discourse. It evaluates MARBERT and XLM-R(oBERTa), with BERT serving as a general-purpose baseline. The models were evaluated on their ability to classify context-sensitive pragmatic functions in code-switched social-media discourse. 11695 unique X posts were collected via Python and utilized for the study. The study employs a quantitative and qualitative NLP approach, following a supervised pipeline. Findings unveil that MARBERT consistently surpasses XLM-R with validation Macro F1 increasing from 0.39 to 0.84 and validation loss decreasing from 0.55 to 0.19. On an independent test set, it achieved 0.96 accuracy, 0.83 macro precision, 0.87 macro recall, and 0.85 Macro F1, while XLM-R achieved 0.92 test accuracy but a substantially lower Macro F1 of 0.52. This was also supported by class-level performance where MARBERT outperforms XLM-R considerably with F1 improvements ranging from +0.33 to +0.60, demonstrating a clear advantage in modeling Arabic digital pragmatics. The study concludes that Transformer performance depends more on DSPP than multilingual coverage alone, as the latter does not guarantee optimal performance on a highly specialized pragmatic classification task.

Multi-agent LLM systems combine multiple inference calls, but prior work often confounds how calls are connected with how they are diversified. We study these factors independently: inference topology and source of inter-agent diversity. In a controlled $2 \times 3$ matrix, we cross parallel aggregation and sequential refinement with stochastic sampling, role prompting, and learned QLoRA specialization, under a fixed three-call budget and output protocol within each backbone. Using Qwen2.5-14B-Instruct and Llama-3.1-8B-Instruct, we evaluate all six configurations on multilingual low-resource emotion detection across nine languages. Parallel learned specialization is strongest on Qwen at 52.83 Macro-F1 and reaches 52.94 on Llama. On Qwen it also exceeds same-backbone zero-shot, few-shot, CoT, and seven-call self-consistency baselines. The preferred topology depends on diversity source: sequential refinement helps stochastic and prompted settings, while the learned Width advantage shrinks from 2.83 points on Qwen to 0.17 on Llama. Depth-wise analysis suggests that later learned specialists can overwrite correct early predictions, although the aggregate effect is backbone-dependent. Overall, how agents are differentiated produces larger performance shifts than topology, which should be evaluated jointly with specialization.

LLM-based sequential recommenders usually cast next-item prediction as text generation, but this interface is poorly matched to full-catalog top-K ranking. We propose TATK, a Triple-Aware framework that couples Top-K Learning (TKL) with Knowledge-Grounded Verification (KGV) for LLM-based sequential recommendation. Top-K Learning combines context-aware metadata-KG prompt grounding with position-aware top-K rewards, aligning training with ranking utility; Knowledge-Grounded Verification then applies structure-aware reranking over the top-M candidates after a single LLM forward pass, using the same metadata-derived item graph. We evaluate TATK on Musical Instruments, CDs and Vinyl, and Video Games from Amazon Reviews 2023 under a matched R2ec-style full-catalog protocol. Experiments use Gemma-2-2B-It and Qwen2.5-3B-Instruct backbones, compare against sequential, generative, KG-augmented, and reasoning-enhanced baselines, and include component, reward-shape, sequence-perturbation, reranking, relation-quality, and candidate-pool diagnostics. TATK improves over the matched R2ec reproduction on all 36 reported metrics. On NDCG@10, it improves Qwen by 8.05%, 4.26%, and 3.78% on the three datasets, and improves Gemma by 27.03%, 10.52%, and 10.23%, while keeping inference within 1.17x of Base RecPO latency. The diagnostics show that structural evidence is most useful for recoverable top-M candidates with reliable KG support, and should be gated when metadata relations are sparse or noisy.

Neyshekar is presented as an open Persian read-speech corpus designed for coverage of both formal and informal language, named entities, and longer utterances. In version 6, 62,279 validated recordings totalling 99.02 hours are provided from 190 contributors, with 34,541 distinct recorded prompts. The prompt pool was assembled from human-written material, contextualised homographs, and reviewed language-model-generated text. Text entries were normalised with the shekar library, which supports both formal and informal Persian, and every submitted recording was reviewed against a common validation rubric. About 24% of released clips are classified as informal by an automatic classifier; these register labels are not human-validated. Item-level rater labels are provided for reproducible agreement estimation, opaque per-clip contributor identifiers make the speaker-disjoint partitioning auditable and support contributor-clustered uncertainty estimates, and a text-disjoint test subset is included for evaluation beyond previously seen prompts. Per-contributor recording load and reference-free signal quality are characterised for every released clip. Corpus characteristics are compared with Persian Common Voice under shared processing. Utility is assessed through two ASR architectures, three optimisation seeds, WER and CER, and independent evaluation on the public PSRB sample. Against duration-matched Common Voice training at approximately 32 hours, in-domain WER is reduced by 9.5 points for Whisper and 11.6 points for XLS-R, and by approximately eight points for both architectures on the independent PSRB sample. Transfer and mixture benefits are not consistently observed across architectures and training budgets. The corpus is released under CC0; code and data are made available through the project repository at https://github.com/amirivojdan/neyshekar.

Rigorous empirical validation of quantum machine learning on natural language tasks remains scarce. We evaluate a 10-qubit hybrid quantum-classical variational circuit (2,148 parameters) for paraphrase detection across three benchmarks: MRPC, Quora Question Pairs (QQP), and adversarial PAWS. On QQP (n = 10 seeds), the circuit achieves 75.53% +/- 0.75%. accuracy, statistically outperforming parameter-matched classical baselines (DeepMLP: p = 0.015, Cohen's d = 1.20; F1: p < 0.001, d = 2.32) and surpassing DistilBERT-4bit with 31,191 fewer parameters. On MRPC the optimal 2-layer variant reaches 92% of BERT-base accuracy at 54,000 lower parameter cost. Circuit depth analysis reveals a dataset-depth scaling effect; entanglement analysis via the Meyer-Wallach measure identifies multi-qubit entanglement as the primary performance driver (r = 0.85 across four variants). Adversarial evaluation on PAWS reveals emergent robustness: 98.2% recall versus 81.6% classical (+16.6 pp, d = 1.24, p < 0.001), without adversarial training. These results constitute the first systematic parameter-matched multi-benchmark empirical validation of hybrid variational circuits for NLP. All results are from classical simulation; hardware validation is future work.

Multi-Hop Knowledge Graph Question Answering (KGQA) tasks require models to assemble relational evidence along paths in a KG to answer natural-language questions. However, existing KGQA systems typically focus on predicting the final answer without explicitly modeling or validating the intermediate reasoning steps, obscuring whether the correct answers arise from faithful multi-hop reasoning. To address this limitation, we re-frame multi-hop KGQA as a question-conditioned graph navigation problem. We refer to this formulation as THESEUS - Traceable Hop-wise Evidence SEarch in a Unified Semantics. In this setting, an agent receives a KG, a question, and a topic entity, and traverses a sequence of relations towards the answer, making the reasoning path explicit. To systematically study this formulation, we provide three key contributions. (i) We augment the existing KINSHIP and MQuAKE resources into navigation-ready KGQA datasets with annotated evidence paths and paraphrased questions. (ii) We design evaluation protocols to measure path fidelity, robustness to linguistic variation, and performance across multi-hop and multi-answer questions. (iii) We adapt established path-based KG completion agents - MINERVA, MultiHopKG, and SQUIRE - to operate on full question embeddings rather than symbolic single-relation queries, enabling their trajectories to be guided by natural-language semantics. Together, these contributions advance KGQA research toward systems where traceability is fundamental: answers are accompanied by explicit reasoning paths whose agreement with reference evidence can be systematically evaluated.

Hallucination in large language models reduces their reliability and slows adoption. Various white-box studies have used internal representations to detect patterns of truthfulness and factuality. A less-studied approach is to identify feed-forward neurons correlated with hallucination. We propose a method to rank feed-forward neurons at the final prompt token using a custom neuron selection dataset. We transfer the selected neuron identities to train hallucination classifiers on other factual question answering datasets. Our work provides empirical evidence that probes trained using the features from the selected neurons perform on par with probes trained on internal states. We also analyze the distribution of selected neurons and the effect of layer depth on detection performance.

Deep research agents answer complex questions through iterative loops of searching, reading, and reasoning. Recent work on reasoning-intensive benchmarks such as BrowseComp-Plus shows that well-configured lexical retrieval can surface high-quality evidence, yet agents may still fail to connect documents carrying evidence to the gold documents. We identify a deep research agent's first retrieval move as an important design decision for this setting. We introduce Question's Gambit, a first-move retrieval module that decomposes the question into a set of clues, reformulates them into complementary searches, consolidates the retrieved results, and reranks the candidate pool before the agent begins its iterative search-and-reasoning process. This produces an opening context designed to support both clue aggregation and final-answer verification. We further evaluate on MultiHop-RAG to test whether these benefits transfer beyond BrowseComp-Plus to a more conventional multi-hop question structure. Experiments on BrowseComp-Plus show that Question's Gambit improves retrieval recall and downstream agent accuracy over strong baselines, improving answer accuracy from 83.1% to 90.5% with gpt-5.5 over Pi-Serini, the strongest reported agentic baseline. Our results confirm that effective agentic deep research depends not only on the tools available inside the loop, but also on the quality of the first move. We published our implementation publicly at https://github.com/radinhamidi/Question-s-Gambit.

Agent benchmarks evaluate policy compliance but assume each policy determines a unique correct action. Natural-language policies can violate this assumption through silence, ambiguity, or contradiction, admitting multiple defensible readings that a single gold trajectory cannot capture. Auditing two $\tau^2$-bench domains, we develop a taxonomy of such policy loopholes and show that affected tasks produce unreliable scores: they lower scores across different models in different ways and make every model less consistent across repeated trials. A cross-domain comparison reveals that exploitability requires both policy ambiguity and tool permissiveness: when policy complexity exceeds what tools can enforce, agents resolve gaps inconsistently and scores become unreliable. Policy specification quality sets the ceiling on evaluation quality. Benchmark developers should audit policies before collecting gold annotations.

Natural language prompts and skills serve as the strategic backbone of LLM-based agents. Recent advances in prompt and skill optimization have achieved notable gains, yet all existing methods optimize a single text template—missing the synergy among multiple complementary strategies. We propose MOSCOPT, a text-native, parameter-free algorithm that jointly optimizes a pool of $N$ skills and a gating skill $G$ that dynamically selects $K$ skills per step. To effectively optimize the skills, we build the EditAdam with internally maintained dual states. Through the three-phase interleaved updates with EditAdam, the system monotonically improves without gradient or parameter tuning. Extensive experiments and detailed ablations across 5 benchmarks and 3 target LLMs demonstrate that MOSCOPT consistently outperforms all baselines, and confirm that both the mixture-of-skills architecture with selective activation and the collective evolution with three-phase interleaving are essential to its superior performance. Code is released https://github.com/zhangzhenyu13/SummerClaw/tree/master/summerclaw/agent_trainer/algorithms/moscopt.

What must a neural system be capable of to implement language? Current research annotates stimuli with linguistic variables and tests which electrodes, voxels, or language-model layers predict neural activity. Yet predictive success leaves mechanisms under-constrained. Here, we show that algebraic properties of language specify invariants that mechanisms must preserve: non-associative hierarchical grouping, commutativity, recursive closure, access to substructures, and structured workspace transitions. We term this the Neural Admissibility Program (NAP). Syntactic structure building is analyzed algebraically, with candidate mechanisms offered for each requirement: content-addressable workspace memory, graph-structured transient dynamics scheduling structure-building operations (e.g. stable heteroclinic channels), and a phase-coupled sealing operation recording grouping. Simulations show that a corrected Marcolli-Berwick entropy-optimized binding gate preserves grouping only within a narrow commitment band. As an alternative, we propose a novel neural binding operation we term 'Meld': two constituent populations converge through shared synapses, integrate sublinearly, and saturate. Meld is, to our knowledge, the closest neurally plausible composition law to syntactic Merge. It preserves every NAP invariant, uses known cortical operations, and recovers hierarchical structure at every tested depth and temperature. It predicts that effective population dimensionality separates alternative bracketings and that the composite depends on constituent disagreement. Importantly, the laws decoding bracketing most accurately are a priori inadmissible, showing that decoding accuracy alone cannot adjudicate between mechanisms. By specifying how neural dynamics can remain faithful to linguistic structure, the NAP changes the criterion by which neural implementations of cognition are evaluated.

Recently, linear attention layers have been increasingly adopted to replace softmax attention at scale for long-context modeling. However, existing context extension approaches typically apply continued pretraining directly without modifying these layers, overlooking the spectral properties of linear attention state dynamics. In this work, we study long-context extension of Gated DeltaNet (GDN) from a spectral perspective of transition matrix and identify two essential factors governing long-range information retrieval: (1) a sufficiently broad slow spectral band aligned with the target dependency length, and (2) the preservation of fast-decaying modes for state clearing and context switching. Based on this observation, we propose SpectralShift, a spectral reparameterization approach for long-context continual pretraining of GDNs. Specifically, SpectralShift reparameterizes the alpha projections initialization to reshape the decay spectrum by enhancing slow propagation capacity, and further introduces a learning-rate scaling for alpha projections to facilitate long-context training. Experiments show that SpectralShift consistently improves long-context capabilities over training, providing an effective and efficient solution for extending context windows of linear attention models. The code has been open-sourced at https://github.com/RUCAIBox/GDN-SpectralShift.

Can financial vision-language models (VLMs) turn chart evidence into reliable action recommendations? Existing hallucination evaluations are mostly claim-centric; they assess whether generated statements are supported, but not whether evidence remains traceable through rationale, confidence, and final action. We introduce E2A-Bench, a 969-query benchmark for financial chart reasoning, constructed from 323 HS300 constituents under three input modalities with deterministic OHLCV-derived evidence anchors. E2A-Bench evaluates grounding, reasoning-action consistency, evidence-confidence calibration, and directional coverage through UCR, RCI, ECI, and NDR, where NDR measures coverage-aware evidence-to-action reliability rather than realized trading performance. Evaluating 20 VLMs reveals three failures hidden by scalar hallucination scores: the lowest-UCR model ranks near the bottom by NDR due to only 6.4% directional coverage; oracle-aided verification reduces unsupported claims but can collapse coverage; and financial fine-tuning amplifies the BUY:SELL ratio by factors of 4.21 to 4.68 across strict base-fine-tuned pairs. These results show that financial VLM evaluation should trace the full evidence-to-action chain rather than rely on a single hallucination score. Code and data: https://github.com/wanng-ide/E2A-Bench

Large language models increasingly analyze computational results and draft manuscripts, making reliable communication as important as correct analysis. Using fixed computational evidence, we tested whether assigning comparisons across modeling choices elsewhere in a research workflow changes manuscript reporting. In constrained sentence-writing tasks, Anthropic's Claude Sonnet 5 often omitted numerical qualifications when detailed comparisons were assigned to a group repository, but retained them more often when the same comparison was assigned to Supporting Information or its own working notes; Claude Opus 5 was less sensitive. These effects did not follow a simple accessibility ordering. A targeted placement rule largely restored sentence-level qualification, whereas a generic accuracy reminder did not. Longer contributions retained numerical qualifications, although some summaries across computational settings were still redirected to the repository. Thus, documenting context within an AI workflow does not ensure its communication where readers encounter the result.

Semantic text watermarks encode signals in meaning rather than surface token choices, offering robustness to paraphrasing and other semantic-preserving edits. Existing semantic watermarking methods are primarily designed for autoregressive language models (ARLMs), where completed candidate units can be generated and scored before generation proceeds. This paradigm does not naturally extend to diffusion language models (DLMs), where semantic units remain incomplete during intermediate denoising steps and tokens may be updated in flexible orders. We propose DenMark, a semantic watermarking framework that injects key-dependent signals directly into the DLM denoising process. DenMark partitions the output into fixed token regions and uses temporary rollouts as semantic lookahead: conditional completions estimate the eventual semantics of an incomplete region, enabling DenMark to select local updates with higher estimated semantic watermark scores. Repeating this procedure across denoising steps progressively accumulates watermark evidence in the final output. For detection, DenMark uses calibrated scanning over candidate unit sizes to remain robust to boundary shifts introduced by semantic attacks. Across four DLMs, three datasets, and four semantic attacks, DenMark gives the strongest overall robustness and achieves the best TPR at each reported operating point in 46 of the 48 backbone-dataset-attack combinations. Overall, DenMark provides an effective semantic watermark for DLMs. Code is available at https://github.com/TianhaoMa5/DenoMark.

Topic models are widely used to analyze public health-related social media short texts, yet their evaluation remains dominated by metrics that focus entirely on generated topics alone. There is a lack of metrics that quantitatively assess whether assigned topics meaningfully represent the corresponding short-text posts. We propose Document-Topic Alignment metrics (DoTA), an assignment-aware evaluation framework comprising metrics that measure semantic alignment between documents (posts) and their assigned topics. We also introduce margin-based and discriminative variants that capture topic assignment confidence and distinguishability. We evaluate DoTA across five topic models on three public health-related social media datasets from X and compare DoTA metrics with conventional topic-based metrics. Results show that DoTA provides complementary evaluation cues and aligns meaningfully with human evaluations. These findings establish the need for assignment-aware evaluation and demonstrate that the addition of DoTA enables a more comprehensive and practically meaningful evaluation for assessing short-text topic modeling performance.