Open-Ended Responses

Graph-based retrieval-augmented generation (RAG) can help answer questions that require information from many documents. However, building a graph often requires many language-model calls during ingestion. It is therefore important to ask whether its quality gains justify the additional cost. We present EffiRAG, a graph-based RAG system designed to reduce this cost. It uses the graph to locate relevant passages and generates answers from the original text. This design preserves source information while keeping graph construction and query processing lightweight. We evaluate EffiRAG on UltraDomain, which contains 120 open-ended questions from four domains. Compared with LightRAG-hybrid, EffiRAG produces the preferred answer on 93 questions. LightRAG is preferred on 7, and the remaining 20 are splits. EffiRAG also reduces total system cost by 57 percent, from USD 0.952 to USD 0.408. The cost includes language-model calls during ingestion and querying. The advantage remains as the corpus grows. At 10 and 20 documents per domain, EffiRAG uses a lightweight, non-LLM filter to skip low-salience chunks. It remains preferred over LightRAG-hybrid. It costs 4.2 times and 4.5 times less, respectively. The comparisons identify different quality-cost trade-offs. Graph-based RAG systems should therefore be evaluated by both answer quality and cost. The results favor graph structure that locates and preserves source evidence.

The development of automated speaking assessment (ASA) is limited by the scarcity of public datasets, with most existing work relying on read-aloud speech, which limits applicability to real-world communication scenarios. In this work, we introduce OpenEnded, a corpus of English practice speech from Mandarin speakers in open-response tasks. Unlike prior open-response datasets that provide only holistic proficiency scores, OpenEnded offers utterance-level assessments of accuracy, fluency, and prosody. Approximately 10,000 utterances are collected and annotated using a hybrid framework: 1,000 are manually labeled via multi-rater scoring with discrepancy resolution to form a high-quality test set, while the remaining are pseudo-labeled by an audio language model (ALM) for training and development sets. We evaluate ALMs and existing ASA models on the OpenEnded test set and introduce VoxPA as an additional baseline. Results show that ALM-generated pseudo-labels improve training over original ALM scoring, while VoxPA achieves the best performance among all baselines.

Blinded physician evaluation has been considered by many to be the gold standard for assessing clinical reasoning in large language models (LLMs). This is difficult to scale; thus, prior studies typically rely on small physician panels, often from a single institution or specialty, which both limits the scientific questions investigated and makes it unclear whether findings would be reproduced with a different set of evaluators. To more rigorously and scalably study clinical reasoning in AI models, here we introduce PrecepTron, an LLM fine-tuned for physician-level evaluation of open-ended responses. PrecepTron was trained using low-rank adaptation (LoRA) of a 32-billion-parameter model on a small number of physician examples. We also release GRAND-ROUNDS, a new large-scale physician-annotated benchmark of 9,217 scores by 11 physicians across seven studies. We show that frontier LLMs in typical "LLM-as-a-judge" approaches often disagree with physicians and with each other, but fine-tuning PrecepTron on a small number of cases enables physician-level consistent scoring across tasks. We use PrecepTron to reproduce headline findings from five influential studies assessing LLMs for clinical care in JAMA, Science, and Nature Medicine without new human grading. Using PrecepTron, we then pose new questions about how LLMs reason in medicine that would have been infeasible with human grading alone, including measuring the diagnostic accuracy of frontier LLMs when clinical cases are provided piecemeal, even token by token. Together, PrecepTron and GRAND-ROUNDS provide a foundation for reproducible, large-scale study of how LLMs reason in medicine. All code, data, and labels are made freely available for researchers.

Maritime Autonomous Surface Ships (MASS) and AI- supported decision assistants are expected to transform maritime operations, but their safe integration depends on how maritime professionals perceive and trust such systems. This paper presents a survey study on maritime stakeholders' attitudes toward an AI-supported assistant in collision-avoidance scenarios. Participants evaluated technology anxiety, trust in automation, and explanation quality using established and adapted questionnaires, complemented by sentiment and thematic analysis of open-ended responses Results indicate a generally positive disposition toward maritime technology, no clear age-related differences in openness, stable trust across scenarios, and more scenario-sensitive, multidimensional explanation ratings. Open responses showed that participants valued support for decision-making, situation awareness, and confidence-building, while raising concerns about AI reliability, over- reliance and loss of expertise. The findings suggest that maritime AI systems should not focus solely on increasing automation or trust, but on supporting calibrated reliance through transparent, reliable, and operationally meaningful design with domain experts in the loop.

Reliable evaluation of open-ended question answering remains a bottleneck for measuring answer correctness of modern LLMs. Unlike multiple-choice tasks, free-form answers may be correct in many surface forms and may fail in qualitatively different ways, including incompleteness, contradiction, overgeneration, and endorsement of false premises. Existing judgment-based and similarity-based metrics often collapse these distinctions. We address this gap with three reusable contributions. First, we introduce a semantic correctness taxonomy that assigns open-ended answers to eight ordered classes, separating verbose-but-correct answers from those contaminated by hallucinated content. Second, we release CAP-Correctness, an 8.8k-example benchmark spanning widely used QA datasets, and CAP-Statements, an 11k-example dataset for converting question-answer pairs into declarative statements for natural language inference (NLI) training and statement-based evaluation. Third, we introduce CAP (Context-Aware Precision), a reference-based metric that scores question-conditioned statements using bidirectional NLI. Under a monotonicity protocol testing whether metrics respect the taxonomy's intended ordering, CAP outperforms established baselines.

Frontier LLMs are increasingly put to use on open-ended complex questions, different in nature from the ones they are typically evaluated on. We dedicate more than 4,000 human expert hours to evaluate a selection of six frontier LLMs on a member of this class of problems: EuroExec, our introduced human expert-based benchmark composed of 413 open-ended long-form European executive tasks authored by 47 vetted domain experts, each question drawn from experience in a real case. Every response is manually evaluated through a multi-attribute rubric, an item-specific checklist of requirements, and a preference rank ordering, extracting an aggregate metric "Solve Rate". The strongest model solves only 56.9% of tasks, while expert-written reference answers judged blindly are solved at near-ceiling levels and are preferred over every model response in 74% of direct rankings, placing frontier generative systems well below the professional standard of work they are already used for. We see that the best way to extract this kind of conclusion is by employing human evaluators, carefully checking their consistency through rigorous statistical analysis, and observe that automatic measurements also fall short when evaluating on this case of real-world open-ended problems with a subjective ground truth.

Prior work on LLM conformity largely measures discrete answer flips under verifiable labels. Open-ended revisions require a different measurement strategy because answer quality is graded, latent, and judged imperfectly. We introduce an experimental protocol implemented across a pooled main peer-condition corpus and separately constructed decomposition corpora, allowing us to separate ordinary re-answering, candidate-content exposure, a bundled peer-presentation residual, and directional judge sensitivity to visible peer context. Across four open-weight generators and three benchmarks, all-wrong peer input produces the lowest-quality revisions in every generator-dataset cell. Blind and informed ratings of identical answers also differ by evaluator: one judge shifts toward the peer-endorsed position, two shift away, one is approximately neutral, and GPT-4o and GPT-5.4-mini audits are likewise non-neutral. Finally, an anchor audit shows that terse correct anchors can be misread often enough to destabilize the latent scale unless calibration is checked explicitly. These results support four conclusions: flip rates are insufficient as a complete measure of open-ended conformity, wrong peers harm open-ended revision, evaluators are not neutral, and anchor calibration is necessary.

LLM agents are increasingly applied to open-ended everyday requests that span work, study, and life. These tasks are long-horizon, cross-environment, and multimodal, forcing the agent to preserve goals and constraints across many steps while navigating heterogeneous tools and attachments. While prior work has addressed individual failure modes such as goals drift, states loss, and context overflow, whether a single harness can manage them jointly and remain effective across backends has received less study. We present OneDayAgent, a long-horizon harness for autonomous agents. OneDayAgent turns an open-ended request into a managed execution process that decomposes tasks into bounded subtasks, maintains execution memory under context pressure, and verifies and repairs the final deliverable. We evaluate OneDayAgent on AgentIF-OneDay across 104 tasks. With the GLM-5.2 backend, OneDayAgent sets a new state of the art with an overall score of 0.821. The same harness runs across five backend LLMs from three model families, indicating the harness generalizes across backends without tuning, even as different models induce distinct execution styles under the same workflow.

AI-generated image forgeries are becoming increasingly realistic and difficult to characterize with fixed manipulation patterns. As generative models continue to evolve, it is impractical to expect a localization model to exhaustively learn all possible forgery appearances from large-scale training data alone. Nevertheless, many AI-generated forgeries still leave subtle forensic traces, although these cues are often weak and unevenly reliable across regions. Therefore, robust localization requires not only extracting informative forensic traces, but also making reliable decisions from incomplete and ambiguous evidence. In this paper, we move beyond static one-shot prediction and reformulate final forgery localization as an adaptive sequential decision-updating process, where the localization map is treated as an intermediate state rather than a fixed output. Rather than producing the final mask via one-shot pixel-wise prediction, our method progressively updates the localization state guided by available evidence, uncertainty, and boundary conditions. Specifically, we first transform mesoscopic traces into compact decision evidence via a lightweight decision evidence projector, and then introduce Evidence-Guided Mamba (EG-Mamba) to perform uncertainty- and boundary-aware state updating. This design allows reliable manipulated and background regions to be preserved, while ambiguous regions are cautiously revised according to the available evidence. Extensive experiments on both conventional and AI-generated manipulation benchmarks validate the effectiveness of the proposed method. Notably, even when trained only on conventional manipulation data, our method brings larger gains on unseen AI-generated forgeries, indicating that progressive decision-updating is especially useful for heterogeneous and hard-to-exhaustively-learn manipulation traces.

Test-time reinforcement learning (TTRL) enables language models to self-evolve at inference time without labeled feedback. Existing methods rely on answer voting and therefore do not extend naturally to open-ended generation, where valid responses cannot be mapped to a shared canonical answer. Without external reward models or stronger judges, adaptation must instead construct reliable rewards from the model's own outputs. We introduce SERPO (Self-Evolving Rubric Policy Optimization), which replaces answer voting with a closed loop that co-evolves response evidence, query-specific rubrics, and policy parameters. Good-Normal-Bad (G-N-B) response evolution organizes maximally separated rollouts into ordered archives; rubric evolution retains criteria that discriminate these archives; probabilistic criterion scoring converts verdict-token likelihoods into reward signals; and policy evolution optimizes the actor with the resulting signals. New actor rollouts then refresh both the archives and rubrics, closing the three-way evolution loop. Across two model configurations, two in-domain benchmarks, and four OOD benchmarks, SERPO improves HealthBench and ResearchQA by up to 20.63 and 20.31 points over the corresponding base models, raises the six-benchmark macro-average by up to 8.06 points, and supports OOD transfer and continued cross-benchmark evolution.

LLM judges are increasingly being used to evaluate open-ended model responses, often in no-reference settings where a ground-truth answer is unavailable. However, can they reliably assess in such evaluation setups? We explore this question in this paper through a two stage pipeline with a) calibration experiments that assess the judge model's knowledge of the task it is evaluating, and b) sensitivity experiments that assess how the judge model's performance is impacted by the presence and positioning of the reference answer in the prompt. Across experiments covering three languages, we show that the judge models we evaluated tend to over-credit incorrect answers in the absence of a reference answer, and adding reference answer information to the prompt flips the judge model's correct/incorrect decisions by as much as 85% in some experimental settings. Comparison with a subset of human annotations shows that these reference-driven changes generally align with human judgments. Our results emphasize the need for calibrating the LLM judges with a sample with reference-aware evaluation before using them in reference-free setups reliably, and our methodology provides a blueprint for researchers and practitioners in doing such calibration of LLM judges for other tasks.

Personalization changes what a model says to a user; we show that it can also change the reasoning trajectory used to justify the response. Modern LLMs personalize interactions by storing user attributes, preferences, and prior context, then injecting this information into future prompts. We study whether such memory reshapes reasoning on open-ended questions where no single ground-truth answer exists. To quantify this effect, we introduce DRIFTLENS, a ground-truth-free framework that maps each expressed reasoning step to a value category and measures divergence between a question's no-memory trajectory and its trajectory under injected user-attribute memory. We first validate that DRIFTLENS distinguishes content-free pragmatic noise from substantive reasoning changes. Across four LLMs and 10 user-attribute categories, including age, occupation, and disability, user-attribute memory induces medium-to-large reasoning drift above each model's pragmatic-noise floor, even when final answers remain fluent, on-topic, and plausible. We then evaluate GRPO- and DPO-based post-training methods for reducing drift. Both reduce drift, but neither uniformly dominates; effects on downstream capability, helpfulness, and instruction following are model-and reward-dependent. These results suggest that memory-induced reasoning drift is a measurable and only partly mitigated failure mode of personalized language models.

Large language models are increasingly used as open-ended search operators in evolutionary optimization. We introduce Evolutionary Feature Engineering (EFE), a framework for using LLM-based evolution to discover preprocessing transformations for structured data. EFE represents transformations as Python programs with a standardized fit/transform interface, allowing them to be inserted directly into existing machine learning pipelines. During evolution, candidate programs are refined using dataset context, summary statistics, and downstream performance feedback on validation set. We instantiate EFE in two settings. For time-series forecasting, EFE-Time learns invertible, dataset-specific normalizations that improve off-the-shelf time-series foundation models. It reduces forecasting errors (MASE, WQL, MAE) 3% or more when averaged across datasets and improvements are as much as 19% on the COVID-Deaths dataset. Notably, these improvements occur with recent TSFMs such as Chronos-2. For tabular prediction, EFE-Tab evolves compact feature programs that add useful interpretable features and remove redundant ones, improving or matching existing LLM-based feature-engineering methods. We found EFE-Tab to be particularly effective on classical decision trees, where small sets of evolved features yield competitive accuracy while preserving interpretability. Overall, EFE demonstrates that LLM-based evolution can improve both accuracy and interpretability when automatically tackling structured data.

Evaluating LLM agents requires dynamic environments that go beyond static reasoning and zero-sum games. Real-world economic interaction is often open-ended and mixed-motive: agents must negotiate, create positive-sum surplus, compete for scarce assets, and plan under delayed returns. We introduce SidConArena, a new benchmark framework for evaluating LLM agents in open-ended, positive-sum bargaining. SidConArena formalizes a multi-player economy as a finite-horizon partially observable stochastic game with three coupled phases: natural-language negotiation with binding trades, deterministic converter-based production, and sealed-bid auctions for long-term assets. The framework combines structured observations, phase-aware agent dispatching, a neural-symbolic action interface, and asynchronous execution, enabling free-form interaction while preserving rule-grounded evaluation. Across homogeneous and heterogeneous tournaments, stronger frontier models achieve higher economic outcomes, yet agents still misvalue resources, bargain passively, and remain limited in long-horizon investment planning.

Although Large Language Models (LLMs) excel in many tasks, their assessment in Portuguese has received less attention, particularly for open-ended, discursive tasks that demand deeper reasoning and generation capabilities. While the original BLUEX benchmark addressed the scarcity of Portuguese evaluation datasets through multiple-choice questions from Brazilian university entrance exams, it did not cover the more challenging second-phase examinations, which require free-form written responses. In this work, we introduce BLUEX v2, a benchmark derived from the second-phase entrance exams of Brazil's two leading universities: UNICAMP (Comvest) and USP (Fuvest), spanning exam years 2022--2025. Our dataset comprises 395 questions unfolding into 919 graded subquestions, with 55.7% of questions containing associated images (represented as context-aware captions during inference to enable evaluation across both vision-capable and text-only models). Each question is annotated with subject area, official reference answers, LLM-generated rubric criteria, and six cognitive capability tags. We evaluate 21 state-of-the-art LLMs using an LLM-as-a-judge protocol. Results reveal a 4.92-point performance spread across models (4.18-9.10 on a 0-10 scale), with Mathematical Reasoning and Image Understanding emerging as the hardest capability dimensions. The evaluation code, model outputs, and dataset are publicly available at https://github.com/TropicAI-Research/BLUEXv2 and on Hugging Face at https://huggingface.co/datasets/Tropic-AI/BLUEX-v2.

There exist numerous tutor training platforms. However, few provide AI-driven training and evaluation for human tutors based on real-life performance. We present an AI-driven system that assesses both open responses during training and authentic real-life tutoring. Unlike platforms that only assess learning through online training or simulations, our system utilizes Generative AI (Gemini-2.5-pro) to analyze transcriptions of authentic tutoring, measuring the transfer of tutor skills to real-life application. Human tutors instructing students remotely in math (N=86) completed six scenario-based lessons, averaging a significant 7.4% learning gain. Using mixed-effects models across 405 session-to-lesson pairs, we found that training performance significantly predicted real-life transcript scores with an effect size of 0.25 SD. Model comparison (AIC/BIC) indicated averaging open response and multiple choice performance during training predicted real-life tutor performance best, although open responses were comparatively more predictive. Exploratory analysis showed that after training, tutors were significantly more likely to encounter pedagogical opportunities to apply their skills (61.1% to 68.9%) and demonstrated higher execution quality within those opportunities (65.5% to 68.1%). Interrupted time series analysis suggested that these tutor improvements were part of a gradual trend over time rather than an immediate intervention effect of training. We illustrate an AI-driven method to link tutor training with real-life assessment. In doing so, we contribute open datasets, AI prompts, and scoring rubrics to support transparency and reproducibility.

For highly capable AI systems to operate safely in dynamic, open-ended environments, they must be able to identify, understand, and respond to moral reasons for action, and constrain their behaviour accordingly. A growing body of research aims to evaluate this capacity -- moral competence -- in today's most capable AI systems, recently reaching broadly pessimistic conclusions. One of the most ambitious such papers collects gold-standard human-authored rubrics for evaluating moral reasoning in 1,000 cases, and benchmarks frontier AI models against those rubrics, with underwhelming results. In this paper, we argue that the MoReBench dataset can be redeployed to give a much more optimistic picture of LLMs' moral reasoning (an essential part of moral competence). We show that if, instead of scoring LLMs' responses to these cases against these rubrics, we instead give the LLMs the same task given to humans -- to generate scoring rubrics for the moral analysis of particular cases -- the rubrics they generate are both better calibrated to the human rubrics than their open-ended responses, and, where they differ, plausibly reflect nothing more than the vast dimensionality of most moral problems, as well as highlighting some human departures from the "rubric for creating rubrics". Taking these points into consideration, the MoReBench dataset suggests that LLMs are significantly more capable at moral reasoning than was previously believed.

A growing body of work points to the great promise of AI systems that can continually expand their capabilities as they operate in an open-ended environment. But yet there is no coherent definition of open-endedness or theory about how an agent ought to explore an open-ended environment. We introduce an information-theoretic definition based on a new concept -- the ${\textit bit-equivalent}$ -- which quantifies the information required to attain each level of expected reward. We consider an environment to be open-ended if an agent can attain linear growth in the bit-equivalent. We establish that classical bandit environments are not open-ended and formulate a bandit environment that is. We also introduce an algorithm that achieves open-ended learning in this environment.

Large Language Models (LLMs) are often fine-tuned through Reinforcement Learning from Human Feedback (RLHF) to align with people's preferences and values. However, this method has known limitations: it aggregates conflicting preferences, often relies on unrepresentative samples, and uses only binary comparisons. Analysing 1,500 open-ended responses from the PRISM dataset across 75 countries, we examine what people actually want from AI systems and reveal concrete failures of current methods. We find that different people want different things: most values are requested by fewer than a quarter of respondents, with truthfulness the sole exception at 49%. Furthermore, the same words hide divergent meanings: when people describe what they mean by "truthfulness", they reveal distinct, potentially incompatible, epistemological bases, as some ask for sourced claims, some for expert opinions, and some even ask for unpopular views. Certain capabilities, namely how human-like a model behaves, and some features, like AI guardrails, are outright controversial, with some desiring them and others rejecting them. We additionally find that people often use contextual distinctions (what AI should do "by default" versus "if requested") that binary comparisons cannot capture. These findings expose fundamental problems in current alignment practices. When 49% request truthfulness but define it differently, this is unlikely to be captured by a single reward model. The persistence of high hallucination rates in well-funded models, despite users' clear demands for accuracy, suggests that current methods fail to identify actual preferences. This paper sheds light on the situated, contested, imperfect signals that are currently being flattened into universal preference models, a practice others have characterised as epistemic violence.

Vision-language models (VLMs), like human learners, are frequently exposed to new visual concepts, but how they map novel visual references to language after exposure remains largely underexplored, particularly when those references contradict prior knowledge from pre-training. To study this, we present the Novel Visual References Dataset (NVRD): 19,176 images spanning 90 visual concepts across different levels of visual novelty, each with up to 20 increasingly perturbed versions of the original object to probe generalization. Unlike prior work on visual augmentations of familiar concepts, NVRD comprises entirely novel, open-ended stimuli constructed from scratch, mirroring how humans encounter genuinely new concepts. We evaluate 3 open- and 2 closed-source models alongside 2,400 human judgments for direct human-model comparison, and find that (i) models struggle to acquire novel concepts in-context when they contradict prior knowledge, and (ii) while models and humans show correlated sensitivity to visual perturbations, models significantly overgeneralize, extending learned labels to stimuli that humans reject. We contribute NVRD as a corpus and benchmark for research on visual concept learning in both humans and machines.

Rubric-based RL is a promising route for extending reinforcement learning beyond verifiable rewards, yet existing methods optimize rubrics while treating the query distribution as fixed. We identify a structural bottleneck: rubric quality is constrained by query structure. Open-ended queries yield vague rubrics; naively narrowing them introduces fabricated references that no model can verify, so all responses fail and training receives no reward signal. We present QUBRIC, a framework that co-designs queries and rubrics. Teacher-derived key points ground the rewriting of open-ended queries into scenario-based, evaluable questions. Contrastive rubric generation then turns teacher-policy gaps into query-level criteria, and learnability filtering retains only informative query-rubric pairs for GRPO training. QUBRIC achieves a +5.5 point gain on ArenaHard over the SFT baseline. Trained only on instruction-following data, it further transfers to three held-out benchmarks spanning legal, moral, and narrative reasoning (+6.3 points on average), with improvements concentrated in reasoning-related dimensions. These results provide evidence that co-designing queries and rubrics can make rubric-based RL a practical complement to RLVR beyond strictly verifiable tasks.

Self-play can train language models without external supervision. However, existing methods require rule-checkable answers, leaving open-ended tasks dependent on curated prompts or frontier-model judges. We introduce SCOPE, a data-free self-play framework for open-ended tasks that co-evolves two policies: a Challenger that generates document-grounded tasks, and a Solver that answers them through multi-turn retrieval. A frozen copy of the initial model serves as the self-judge, which writes task-specific rubrics from the source document and grades Solver responses against them. Across three 7-8B instruction-tuned models (Qwen2.5, Qwen3, OLMo-3), SCOPE improves open-ended performance by up to +10.4 points on eight benchmarks and matches or exceeds GRPO_data trained on ~9K curated prompts. Although trained only on open-ended tasks, SCOPE also improves held-out short-form QA by up to +13.8 points on seven held-out benchmarks, surpassing GRPO_data on all three models. Ablations show that co-evolving the Challenger is necessary to keep tasks near the Solver's frontier, that gains arise from improvements in both retrieval and synthesis with the relative contribution varying by task, and that rubric generation quality is the bottleneck for self-judging.

Post-training large language models with reinforcement learning is bottlenecked by the reward signal. Existing approaches require either ground-truth verifiable rewards, restricting training to domains with automatic correctness checks (e.g., mathematics, code execution), or human preference labels, which are expensive to collect and prone to reward hacking. Recent label-free methods replace ground-truth verifiers with self-referential signals like majority voting or token entropy over a model's own outputs, but risk reinforcing a model's own errors. In this work we propose Cross-Model Entropy (CME), the mean log-likelihood of a generator's response under a separate verifier model, as a label-free reward signal for RL post-training. CME is continuous, training-free, and grounded in the principle that responses a verifier finds unsurprising are likely correct or high quality. Because the verifier is independent of the generator, the signal cannot be gamed through self-consistency. We integrate CME into GRPO with no other changes to the training loop, extending label-free RL to open-ended instruction following -- a regime where self-referential signals are inapplicable or poorly suited. On open-ended instruction following (UltraFeedback prompts, evaluated on AlpacaEval 2.0), CME rewards beat the untrained base in head-to-head LLM-as-Judge comparisons across four model families (Qwen, Llama, Gemma, OLMo) and three training regimes (pretrained, SFT, and instruction-tuned), with tie-adjusted win rates ranging from 52.5% to 71.4%. Code will be released upon publication.

Reliable set-valued prediction provides a principled way to mitigate hallucinations in open-ended question answering (QA), yet existing conformal approaches typically rely on a fragile premise: finite sampling must already produce at least one admissible candidate, or calibration examples violating this condition are discarded. In this paper, we introduce MiRD, a two-stage framework that decomposes overall miscoverage into sampling failure and conditional selection failure. In Stage I, MiRD establishes an expectation-level marginal upper bound on the probability that finite sampling produces no admissible answer under a fixed budget. In Stage II, conditioned on sampling success, MiRD calibrates a conformal selection threshold using admission-correlated nonconformity scores defined over the full calibration set, thereby preserving calibration-set integrity. Across three open-ended QA datasets and eight models, MiRD controls sampling risk, conditional selection risk, and overall miscoverage, while yielding tighter first-stage bounds than PAC-style alternatives and more adaptive prediction sets than successful-only calibration.

In satellite applications, user queries often take the form of open-ended natural language, extending beyond a fixed set of predefined categories. This open-vocabulary nature poses significant challenges for retrieving relevant image tiles, as the retrieval system must generalize to a wide range of unseen objects and concepts. While vision-language models (VLMs) such as CLIP are widely used for text-image retrieval, even fine-tuned variants often struggle to accurately align such queries with satellite imagery. To address this, we propose Open-SAT, a training-free query embedding refinement algorithm that operates at inference time to improve alignment between user queries and satellite image content. Open-SAT uses VLMs to compute embeddings for image tiles, which are stored in a vector database for efficient retrieval. At query time, it leverages Large Language Models (LLMs) to refine the text embeddings by incorporating contextual information about objects of interest and their surroundings. A threshold-free retrieval mechanism further enhances accuracy and efficiency. Experimental results in three public benchmarks demonstrate that Open-SAT improves the F1 score by up to 16.04%, while retrieving a comparable number of image tiles. These results demonstrate the effectiveness of Open-SAT in open-vocabulary satellite image retrieval, leveraging LLM guidance without the need for additional training or supervision.

Purpose: Vision-language models (VLMs) have shown promising performance in surgical visual question answering (VQA). However, existing surgical VQA datasets often contain linguistic shortcuts, where question phrasing implicitly constrains the answer space. It remains unclear whether reported performance reflects visual understanding or reliance on such linguistic shortcuts. Methods: We introduce SurgCheck, a diagnostic benchmark for quantifying linguistic shortcut reliance in surgical VQA. SurgCheck employs a paired-question design in which each surgical frame is associated with an original question containing entity names and a less-biased counterpart that removes these names while preserving identical visual content and ground-truth answers. The resulting performance gap provides a diagnostic signal of shortcut reliance. To ensure that the less-biased question remains well-defined even without entity names, four grounding cues are incorporated: bounding box, arrow, spatial position, and periphrasis. We evaluate both general-purpose and surgical-specific VLMs under zero-shot and fine-tuned settings on SurgCheck. To evaluate open-ended zero-shot responses, we introduce an LLM-as-a-judge evaluation protocol. Results: Using SurgCheck, we observe consistent performance degradation on less-biased questions across five VLMs, despite identical visual inputs. Text-only ablation reveals minimal performance drops for action and target prediction, indicating that action and target prediction is largely driven by linguistic shortcuts rather than visual reasoning. Conclusion: SurgCheck provides a controlled diagnostic framework that exposes failure modes masked by linguistic bias in existing surgical VQA benchmarks. Our findings demonstrate that strong benchmark performance does not necessarily imply faithful visual understanding, underscoring the need for bias-aware evaluation in surgical VQA.

Urban transportation systems face growing safety challenges that require scalable intelligence for emerging smart mobility infrastructures. While recent advances in foundation models and large-scale multimodal datasets have strengthened perception and reasoning in intelligent transportation systems (ITS), existing research remains largely centered on microscopic autonomous driving (AD), with limited attention to city-scale traffic analysis. In particular, open-ended safety-oriented visual question answering (VQA) and corresponding foundation models for reasoning over heterogeneous roadside camera observations remain underexplored. To address this gap, we introduce the Land Transportation Dataset (LTD), a large-scale open-source vision-language dataset for open-ended reasoning in urban traffic environments. LTD contains 11.6K high-quality VQA pairs collected from heterogeneous roadside cameras, spanning diverse road geometries, traffic participants, illumination conditions, and adverse weather. The dataset integrates three complementary tasks: fine-grained multi-object grounding, multi-image camera selection, and multi-image risk analysis, requiring joint reasoning over minimally correlated views to infer hazardous objects, contributing factors, and risky road directions. To ensure annotation fidelity, we combine multi-model vision-language generation with cross-validation and human-in-the-loop refinement. Building upon LTD, we further propose UniVLT, a transportation foundation model trained via curriculum-based knowledge transfer to unify microscopic AD reasoning and macroscopic traffic analysis within a single architecture. Extensive experiments on LTD and multiple AD benchmarks demonstrate that UniVLT achieves SOTA performance on open-ended reasoning tasks across diverse domains, while exposing limitations of existing foundation models in complex multi-view traffic scenarios.

Answering open-ended questions remains challenging for AI systems because it requires synthesis, judgment, and exploration beyond factual retrieval, and users often refine answers through multiple iterations rather than accepting a single response. Existing QA benchmarks do not explicitly support this refinement process. To address this gap, we introduce a new task, document-grounded related insight generation, where the goal is to generate additional insights from a document collection that help improve, extend, or rethink an initial answer to an open-ended question, ultimately supporting richer user interaction and a better overall question answering experience. We curate and release SCOpE-QA (Scientific Collections for Open-Ended QA), a dataset of 3,000 open-ended questions across 20 research collections. We present InsightGen, a two-stage approach that first constructs a thematic representation of the document collection using clustering, and then selects related context based on neighborhood selection from the thematic graph to generate diverse and relevant insights using LLMs. Extensive evaluation on 3,000 questions using two generation models and two evaluation settings shows that InsightGen consistently produces useful, relevant, and actionable insights, establishing a strong baseline for this new task.

Activation-based steering enables inference-time personalization of large language models, but its effects in educational applications are not well understood. We study activation-based persona vectors representing seven character traits in short-answer generation and automated scoring on the ASAP-SAS benchmark, across three language models spanning dense and mixture-of-experts architectures. Persona steering lowers answer quality overall, with much larger effects on open-ended English Language Arts (ELA) prompts than on factual science prompts. Interpretive and argumentative tasks are particularly sensitive, showing up to 11$\times$ larger degradation. On the scoring side, we observe predictable valence-aligned calibration shifts: evil'' and impolite'' scorers grade more harshly, while good'' and optimistic'' scorers grade more leniently. ELA tasks are 2.5-3$\times$ more susceptible to scorer personalization than science tasks, and the mixture-of-experts model shows roughly 6$\times$ larger calibration shifts than the dense models. To our knowledge, this is the first study to systematically examine the effects of activation-steered persona traits in educational generation and scoring. Our findings highlight the need for task- and architecture-aware calibration when deploying personalized models in educational settings.

Autoregressive (AR) video diffusion models in the Self Forcing family generate a video chunk by chunk, each conditioned on the ones before it, so they can keep rolling out with no preset length. In practice they are trained on short, fixed clips (e.g., 5s). Once a rollout passes this training horizon, the cache fills with self-generated chunks the model never saw in training, and quality degrades: colors oversaturate, identities drift, structures collapse, and motion fades. We frame this failure as a cache-behavior mismatch, a domain shift between the within-horizon caches seen in training and those produced by open-ended testing, and address it with Rolling Sink, a training-free inference-time cache policy. Within the base model's fixed cache budget, Rolling Sink holds the cache close to its stable within-horizon behavior: it restores low-drift content, sliding temporal indices, and evolving semantics, and keeps a small recent context. Attention sink and temporal re-indexing are existing mechanisms. Our contribution is the mismatch formulation, the joint bounded-cache policy, and the Rolling Semantics schedule. On two 5s-trained bases, Self Forcing and Causal Forcing, Rolling Sink attains the best averaged rank on VBench-Long at both 1- and 5-minute rollouts under the same inference setting, at about 0.56% throughput change, and holds GPU memory constant. It also sustains a 30-minute qualitative rollout, 360 times its training horizon. We validate at 1- and 5-minute rollouts and do not claim guaranteed quality at arbitrary duration. Project page: https://rolling-sink.github.io/

Reliable assessment of the abilities of large audio language models (LALMs) is essential to advancing the state of the art. As benchmarks rapidly evolve to incorporate complex reasoning and subjective tasks, they increasingly necessitate open-ended responses from LALMs. We present Open-ended Response Correctness Assessment (ORCA) -- a reliable and lightweight model-based approach for answer correctness and disagreement modeling. We employ a three-stage annotation pipeline combining human judgment, structured feedback, and human-AI correction, yielding 9,663 annotations across 3,699 question-answer pairs from 15 LALMs on three audio understanding and reasoning benchmarks (achieving a Krippendorff's alpha of 0.82). Our experiments employing curriculum learning show that ORCA models achieve a Spearman correlation of 0.91 with average human correctness ratings on seen benchmarks and generalize to unseen benchmarks with a score of 0.85, outperforming several LLM judge baselines including Gemini 2.5 Flash. Furthermore, we demonstrate that ORCA's predicted variance correlates strongly with human disagreement, allowing it to effectively identify problematic benchmark items.

Qualitative content analysis of open-ended survey responses is a commonly used research method in science education. However, traditional coding approaches are often time-consuming and prone to inconsistency, especially when applied to large datasets. Existing solutions from Natural Language Processing such as supervised classifiers, topic modeling techniques, and generative large language models have limited applicability in analysis of open-ended survey responses, since they demand extensive labeled data, disrupt established qualitative workflows, and/or yield variable results. In this paper, we introduce a text embedding-based classification framework called Deductive Semantic Content Analysis (DeSCA) that requires only a handful of examples per category to run, is transparent and replicable, and fits well with standard qualitative workflows. When benchmarked against human analysis of a physics education survey consisting of 2899 open-ended responses, the method described by our framework achieves high agreement with expert human coders across ten embeddings models on a simulated exhaustive coding task, using approximately 1-2% of the total dataset for training. The method achieves lower agreement on a complete selective coding task; this performance, however, improves with fine-tuning of the text embedding model, which can be done with a small amount of additional data. We unpack these results in terms of the theoretical assumptions of text embeddings, and further demonstrate how embeddings can be used to audit previously-analyzed datasets for coding consistency. These findings demonstrate that text embedding-assisted coding can flexibly scale to thousands of responses without sacrificing interpretability, opening avenues for deductive qualitative analysis at scale.