{"total":20,"items":[{"citing_arxiv_id":"2605.17946","ref_index":66,"ref_count":2,"confidence":0.98,"is_internal_anchor":true,"paper_title":"SVFSearch: A Multimodal Knowledge-Intensive Benchmark for Short-Video Frame Search in the Gaming Vertical Domain","primary_cat":"cs.AI","submitted_at":"2026-05-18T07:03:48+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"SVFSearch is the first open benchmark for short-video frame search in the Chinese gaming domain, providing a frozen retrieval environment and showing performance gaps of 13-29 points between direct QA models, practical agents, and oracle knowledge.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.12497","ref_index":14,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"From Web to Pixels: Bringing Agentic Search into Visual Perception","primary_cat":"cs.CV","submitted_at":"2026-05-12T17:59:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"WebEye benchmark and Pixel-Searcher agent enable visual perception tasks by using web search to resolve object identities before precise localization or answering.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"InProceedings of the IEEE/cvf conference on computer vision and pattern recognition, pages 3195-3204, 2019. [13] Dongzhi Jiang, Renrui Zhang, Ziyu Guo, Yanmin Wu, Jiayi Lei, Pengshuo Qiu, Pan Lu, Zehui Chen, Chaoyou Fu, Guanglu Song, et al. Mmsearch: Benchmarking the potential of large models as multi-modal search engines. arXiv preprint arXiv:2409.12959, 2024. [14] Jinming Wu, Zihao Deng, Wei Li, Yiding Liu, Bo You, Bo Li, Zejun Ma, and Ziwei Liu. Mmsearch-r1: Incentivizing lmms to search.arXiv preprint arXiv:2506.20670, 2025. [15] Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, Noveen Sachdeva, Inderjit Dhillon, Marcel Blistein, Ori Ram, Dan Zhang, Evan Rosen, et al. Gemini 2.5: Pushing the frontier with advanced reasoning,"},{"citing_arxiv_id":"2605.10832","ref_index":19,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents","primary_cat":"cs.CL","submitted_at":"2026-05-11T16:49:36+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A new image-bank harness and closed-loop on-policy data evolution method raises multimodal agent performance on visual search benchmarks from 24.9% to 39.0% for an 8B model and from 30.6% to 41.5% for a 30B model.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.09934","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"TRACER: Verifiable Generative Provenance for Multimodal Tool-Using Agents","primary_cat":"cs.CL","submitted_at":"2026-05-11T03:32:55+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"TRACER attaches verifiable sentence-level provenance records to multimodal agent outputs using tool-turn alignment and semantic relations, yielding 78.23% answer accuracy and fewer tool calls than baselines on TRACE-Bench.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"a benchmark for provenance-aware multimodal tool reasoning, and show that TRACER improves answer accuracy, provenance faithfulness, and tool-use efficiency. 2 Related Work Multimodal tool-using agents.Recent multimodal agents move beyond single-turn perception toward long-horizon interaction, where models iteratively search, inspect, calculate, and integrate information across tools and modalities. Search-R1 [ 4] and MMSearch-R1 [5] integrate external search into model reasoning, first in text-dominant settings and then in multimodal search over images and text. WebWatcher [ 6] and Vision-DeepResearch [7] study deep-research-style interaction in visually grounded environments. DeepEyesV2 [ 8], Agent0-VL [ 9], and ReAgent-V [ 10] further improve tool-mediated reasoning, training stability, and video-oriented interaction."},{"citing_arxiv_id":"2605.09271","ref_index":115,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Shaping Schema via Language Representation as the Next Frontier for LLM Intelligence Expanding","primary_cat":"cs.AI","submitted_at":"2026-05-10T02:42:29+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"Advanced language representations shape LLMs' schemas to improve knowledge activation and problem-solving.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Chameleon: Plug-and-play compositional reasoning with large language models.Advances in Neural Information Processing Systems, 36:43447-43478, 2023. [114] Shi Yu, Chaoyue Tang, Bokai Xu, Junbo Cui, Junhao Ran, Yukun Yan, Zhenghao Liu, Shuo Wang, Xu Han, Zhiyuan Liu, et al. Visrag: Vision-based retrieval-augmented generation on multi-modality documents.arXiv preprint arXiv:2410.10594, 2024. [115] Jinming Wu, Zihao Deng, Wei Li, Yiding Liu, Bo You, Bo Li, Zejun Ma, and Ziwei Liu. Mmsearch-r1: Incentivizing lmms to search.arXiv preprint arXiv:2506.20670, 2025. [116] Luyu Gao, Aman Madaan, Shuyan Zhou, Uri Alon, Pengfei Liu, Yiming Yang, Jamie Callan, and Graham Neubig. Pal: Program-aided language models. InInternational Conference on Machine Learning, pages 10764-10799."},{"citing_arxiv_id":"2605.07177","ref_index":39,"ref_count":2,"confidence":0.9,"is_internal_anchor":true,"paper_title":"HyperEyes: Dual-Grained Efficiency-Aware Reinforcement Learning for Parallel Multimodal Search Agents","primary_cat":"cs.LG","submitted_at":"2026-05-08T03:16:08+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"HyperEyes presents a parallel multimodal search agent using dual-grained efficiency-aware RL with a new TRACE reward and IMEB benchmark, claiming 9.9% higher accuracy and 5.3x fewer tool calls than prior open-source agents.","context_count":1,"top_context_role":"dataset","top_context_polarity":"use_dataset","context_text":"and enable efficiency-aware optimization, we design a comprehensive three-stage data curation pipeline (illustrated in Fig. 2). First, we compile a diverse pool of tasks by aggregating public 4 Table 1: Composition of the training dataset. Data Source # QA Pairs # SFT # 30B RL # 235B RL Public Benchmarks LiveVQA [7] 100k 13.5k 3k 5k REDSearch [5] 10k 2k 0.5k 0.5k InfoSeek [39] 41k 3k - - iNaturalist [34] 75k 2.5k 2k 3k Google-Landmark [37] 12k 1k - - DeepDive [19] 3k 1.5k - - Ours Internal Human Annotations 5k 0.5k - - Visual Multi-Entity (Sec.3.2.1) 20k 5k 0.5k 0.5k Textual Multi-Constraint (Sec.3.2.1) 5k 1k - - Total 271k 30k 6k 9k datasets and synthesizing novel multi-entity queries (Sec. 3.2.1). Second, we construct a high-quality"},{"citing_arxiv_id":"2605.02378","ref_index":45,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Enhancing Multimodal In-Context Learning via Inductive-Deductive Reasoning","primary_cat":"cs.CV","submitted_at":"2026-05-04T09:18:19+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A framework with similarity-based visual token compression, dynamic attention rebalancing, and explicit inductive-deductive chain-of-thought improves multimodal ICL performance across eight benchmarks for open-source VLMs.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Advances in neural information processing systems, 35:24824-24837, 2022. [43] Noam Wies, Yoav Levine, and Amnon Shashua. The learnability of in-context learning. Advances in Neural Information Processing Systems, 36:36637-36651, 2023. [44] Jinming Wu, Zihao Deng, Wei Li, Yiding Liu, Bo You, Bo Li, Zejun Ma, and Ziwei Liu. Mmsearch-r1: Incentivizing lmms to search.arXiv preprint arXiv:2506.20670, 2025. 12 [45] Youze Xue, Dian Li, and Gang Liu. Improve multi-modal embedding learning via explicit hard negative gradient amplifying.arXiv preprint arXiv:2506.02020, 2025. [46] Yi Yang, Xiaoxuan He, Hongkun Pan, Xiyan Jiang, Yan Deng, Xingtao Yang, Haoyu Lu, Dacheng Yin, Fengyun Rao, Minfeng Zhu, et al. R1-onevision: Advancing generalized multimodal reasoning through cross-modal formalization."},{"citing_arxiv_id":"2604.20486","ref_index":30,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"ProMMSearchAgent: A Generalizable Multimodal Search Agent Trained with Process-Oriented Rewards","primary_cat":"cs.CV","submitted_at":"2026-04-22T12:20:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A sandbox-trained multimodal search agent with process-oriented rewards transfers zero-shot to real Google Search and outperforms prior methods on FVQA, InfoSeek, and MMSearch.","context_count":1,"top_context_role":"baseline","top_context_polarity":"baseline","context_text":"by the extreme sparsity of effective supervision signals. Since standard reinforce- ment learning (RL) approaches rely predominantly on final outcome rewards, models often develop a degenerate strategy: indiscriminately invoking search tools to maximize the chance of stumbling upon the right answer. To prevent models from over-relying on external tools, existing methods like MMSearch- R1 [30] introduce a heuristic \"tool penalty\" to discourage tool calls. However, our empirical analysis reveals that this introduces a criticalreward-action mis- match: for visually ambiguous samples that inherently require external search, the agent is penalized for taking the correct action. This paradox entangles answer correctness with proper tool-use strategy, leaving the agent unable to"},{"citing_arxiv_id":"2604.20146","ref_index":41,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"SAKE: Self-aware Knowledge Exploitation-Exploration for Grounded Multimodal Named Entity Recognition","primary_cat":"cs.IR","submitted_at":"2026-04-22T03:17:36+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"SAKE is an agentic framework for GMNER that uses uncertainty-based self-awareness and reinforcement learning to balance internal knowledge exploitation with adaptive external exploration.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.19264","ref_index":19,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"DR-MMSearchAgent: Deepening Reasoning in Multimodal Search Agents","primary_cat":"cs.CV","submitted_at":"2026-04-21T09:28:34+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"DR-MMSearchAgent derives batch-wide trajectory advantages and uses differentiated Gaussian rewards to prevent premature collapse in multimodal agents, outperforming MMSearch-R1 by 8.4% on FVQA-test.","context_count":1,"top_context_role":"background","top_context_polarity":"unclear","context_text":"=z i,t.(18) From this, it can be seen that the normalization remains unchanged before and after, so for the subsequent steps, we obtainW ′ i =W i. Theorem A.2.Scale invariance of injected advantage. Un- der any affine transform R′(τ) =αR(τ) with α >0 , the A(τ) and the SPAI score W(τ) are invariant, hence the injected advantage A′(τ) =A(τ) +A(τ)·W i.(19) is invariant as well. Remark.In practice it is common to compute a numerically-stabilized score F(τ) = D−(τ) D+(τ)+D −(τ)+ϵ , with a small ϵ >0 to avoid division-by-zero. In practice, this occurs extremely rarely, so the impact can be considered negligible. A.3. Discriminability via Distributional Sparsity Let τa and τb be two trajectories yielding identical scalar"},{"citing_arxiv_id":"2604.14029","ref_index":49,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"POINTS-Seeker: Towards Training a Multimodal Agentic Search Model from Scratch","primary_cat":"cs.CV","submitted_at":"2026-04-15T16:09:37+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"POINTS-Seeker-8B is an 8B multimodal model trained from scratch for agentic search that uses seeding and visual-space history folding to outperform prior models on six visual reasoning benchmarks.","context_count":1,"top_context_role":"dataset","top_context_polarity":"use_dataset","context_text":"ios while preserving reasoning efficacy through history-aware selective rendering. structured tool-use trajectories. Our approach consists of two primary phases: vision-oriented data construction and trajectory generation with filtering. VQA data construction.We curate high-quality VQA data by leveraging search-oriented VQA datasets (i.e., LiveVQA [11] and FVQA [49]) or converting complex multi-hop QA datasets [50]. Our automated conversion pipeline involves the following three key steps: (i)Entity extraction & selection:We use Claude-4-5-Sonnet to identify visually representable entities from questions. (ii) Query de-biasing:To ensure retrieval accuracy, we generate context-aware search queries (e.g., appending \"New South Wales politician\" to \"Mike Baird\") to"},{"citing_arxiv_id":"2604.12890","ref_index":61,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Towards Long-horizon Agentic Multimodal Search","primary_cat":"cs.CV","submitted_at":"2026-04-14T15:40:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"LMM-Searcher uses file-based visual UIDs and a fetch tool plus 12K synthesized trajectories to fine-tune a multimodal agent that scales to 100-turn horizons and reaches SOTA among open-source models on MM-BrowseComp and MMSearch-Plus.","context_count":1,"top_context_role":"baseline","top_context_polarity":"baseline","context_text":"4 11.0 - 69.7 w. Our Framework35.1 46.7 58.0 73.2 Qwen3-VL-30B- A3B-Thinking Direct Answer 7.1 2.7 13.0 17.7 w. Previous Framework10.713.6 - 53.2 w. Our Framework 9.814.4 16.0 62.0 Table 3: Performance comparison among different frameworks. • Multimodal Search Agents. We compare against existing open-source multimodal agents, in- cluding MMSearch-R1 [61], WebWatcher [6], DeepEyesV2 [21], Vision-DeepResearch [7], and REDSearcher-MM [8]. These methods typically combine perception, reasoning, and external search to address complex multimodal queries. Implementation Details.We build our framework based on MiroFlow [ 62], and utilize it for both trajectory rollout and answer verification. We utilize LLaMA-Factory [63] as the training framework."},{"citing_arxiv_id":"2604.09508","ref_index":27,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"VISOR: Agentic Visual Retrieval-Augmented Generation via Iterative Search and Over-horizon Reasoning","primary_cat":"cs.CV","submitted_at":"2026-04-10T17:25:34+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"VISOR is a unified agentic VRAG framework with Evidence Space structuring, visual action evaluation/correction, and dynamic sliding-window trajectories trained via GRPO-based RL that achieves SOTA performance on long-horizon visual reasoning benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.08545","ref_index":41,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Act Wisely: Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Models","primary_cat":"cs.CV","submitted_at":"2026-04-09T17:59:57+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"HDPO reframes tool efficiency as a conditional objective within accurate trajectories, enabling Metis to reduce tool invocations by orders of magnitude while raising reasoning accuracy.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Charxiv: Charting gaps in realistic chart understanding in multimodal llms.Advances in Neural Information Processing Systems, 37:113569-113697, 2024. [40] Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824-24837, 2022. [41] Jinming Wu, Zihao Deng, Wei Li, Yiding Liu, Bo You, Bo Li, Zejun Ma, and Ziwei Liu. Mmsearch-r1: Incentivizing lmms to search.arXiv preprint arXiv:2506.20670, 2025. [42] Penghao Wu and Saining Xie. V?: Guided visual search as a core mechanism in multimodal llms. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13084-"},{"citing_arxiv_id":"2604.06777","ref_index":64,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Walk the Talk: Bridging the Reasoning-Action Gap for Thinking with Images via Multimodal Agentic Policy Optimization","primary_cat":"cs.CV","submitted_at":"2026-04-08T07:48:07+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MAPO improves multimodal chain-of-thought reasoning by requiring explicit textual descriptions of visual tool results and using a novel advantage estimator that combines semantic alignment with task rewards.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"[62] Alex Su, Haozhe Wang, Weiming Ren, Fangzhen Lin, and Wenhu Chen. Pixel reasoner: Incentivizing pixel- space reasoning with curiosity-driven reinforcement learning.arXiv preprint arXiv:2505.15966, 2025. [63] Xinyu Huang, Yuhao Dong, Weiwei Tian, Bo Li, Rui Feng, and Ziwei Liu. High-resolution visual reasoning via multi-turn grounding-based reinforcement learning.arXiv preprint arXiv:2507.05920, 2025. [64] Jinming Wu, Zihao Deng, Wei Li, Yiding Liu, Bo You, Bo Li, Zejun Ma, and Ziwei Liu. Mmsearch-r1: Incen- tivizing lmms to search.arXiv preprint arXiv:2506.20670, 2025. [65] Senqiao Yang, Junyi Li, Xin Lai, Bei Yu, Hengshuang Zhao, and Jiaya Jia. Visionthink: Smart and efficient vision language model via reinforcement learning.arXiv preprint arXiv:2507."},{"citing_arxiv_id":"2604.04500","ref_index":75,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Saliency-R1: Enforcing Interpretable and Faithful Vision-language Reasoning via Saliency-map Alignment Reward","primary_cat":"cs.CV","submitted_at":"2026-04-06T07:51:59+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Saliency-R1 uses a novel saliency map technique and GRPO with human bounding-box overlap as reward to improve VLM reasoning faithfulness and interpretability.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"InProceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 24-25, 2020. 7 [74] Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large lan- guage models.Advances in neural information processing systems, 35:24824-24837, 2022. 1 [75] Jinming Wu, Zihao Deng, Wei Li, Yiding Liu, Bo You, Bo Li, Zejun Ma, and Ziwei Liu. Mmsearch-r1: Incentivizing lmms to search.arXiv preprint arXiv:2506.20670, 2025. 2 [76] Cheng Xia, Manxi Lin, Jiexiang Tan, Xiaoxiong Du, Yang Qiu, Junjun Zheng, Xiangheng Kong, Yuning Jiang, and Bo Zheng. Mirage: Towards ai-generated image detection in the wild.arXiv preprint arXiv:2508."},{"citing_arxiv_id":"2602.22683","ref_index":50,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"SUPERGLASSES: Benchmarking Vision Language Models as Intelligent Agents for AI Smart Glasses","primary_cat":"cs.CV","submitted_at":"2026-02-26T06:55:48+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"SUPERGLASSES is the first VQA benchmark built from actual smart glasses data, and SUPERLENS is an agent using automatic object detection, query decoupling, and multimodal search that outperforms GPT-4o by 2.19% on it.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2511.05271","ref_index":53,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"DeepEyesV2: Toward Agentic Multimodal Model","primary_cat":"cs.CV","submitted_at":"2025-11-07T14:31:20+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"DeepEyesV2 uses a two-stage cold-start plus reinforcement learning pipeline to produce an agentic multimodal model that adaptively invokes tools and outperforms direct RL on real-world reasoning benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2509.07969","ref_index":40,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Mini-o3: Scaling Up Reasoning Patterns and Interaction Turns for Visual Search","primary_cat":"cs.CV","submitted_at":"2025-09-09T17:54:21+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Mini-o3 scales visual search reasoning to tens of interaction turns via a new probe dataset, iterative trajectory collection, and over-turn masking in RL, claiming SOTA performance while training only up to six turns.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2505.22095","ref_index":37,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Mixture-of-Retrieval Experts for Reasoning-Guided Multimodal Knowledge Exploitation","primary_cat":"cs.CL","submitted_at":"2025-05-28T08:17:57+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MoRE enables MLLMs to dynamically coordinate heterogeneous retrieval experts via Step-GRPO training, yielding over 7% average gains on open-domain QA benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}