Total Recall QA: A Verifiable Evaluation Suite for Deep Research Agents

Deep research agents have emerged as LLM-based systems designed to perform multi-step information seeking and reasoning over large, open-domain sources to answer complex questions by synthesizing information from multiple information sources. Given the complexity of the task and despite various recent efforts, evaluation of deep research agents remains fundamentally challenging. This paper identifies a list of requirements and optional properties for evaluating deep research agents. We observe that existing benchmarks do not satisfy all identified requirements. Inspired by prior research on TREC Total Recall Tracks, we introduce the task of Total Recall Question Answering and develop a framework for deep research agents evaluation that satisfies the identified criteria. Our framework constructs single-answer, total recall queries with precise evaluation and relevance judgments derived from a structured knowledge base paired with a text corpus, enabling large-scale data construction. Using this framework, we build TRQA, a deep research benchmark constructed from Wikidata-Wikipedia as a real-world source and a synthetically generated e-commerce knowledge base and corpus to mitigate the effects of data contamination. We benchmark the collection with representative retriever and deep research models and establish baseline retrieval and end-to-end results for future comparative evaluation.

Scaling Laws for Reranking in Information Retrieval

Scaling laws have been observed across a wide range of tasks, such as natural language generation and dense retrieval, where performance follows predictable patterns as model size, data, and compute grow. However, these scaling laws are insufficient for understanding the scaling behavior of multi-stage retrieval systems, which typically include a reranking stage. In large-scale multi-stage retrieval systems, reranking is the final and most influential step before presenting a ranked list of items to the end user. In this work, we present the first systematic study of scaling laws for rerankers by analyzing performance across model sizes and data budgets for three popular paradigms: pointwise, pairwise, and listwise reranking. Using a detailed case study with cross-encoder rerankers, we demonstrate that performance follows a predictable power law. This regularity allows us to accurately forecast the performance of larger models for some metrics more than others using smaller-scale experiments, offering a robust methodology for saving significant computational resources. For example, we accurately estimate the NDCG of a 1B-parameter model by training and evaluating only smaller models (up to 400M parameters), in both in-domain as well as out-of-domain settings. Our experiments encompass span several loss functions, models and metrics and demonstrate that downstream metrics like NDCG, MAP (Mean Avg Precision) show reliable scaling behavior and can be forecasted accurately at scale, while highlighting the limitations of metrics like Contrastive Entropy and MRR (Mean Reciprocal Rank) which do not follow predictable scaling behavior in all instances. Our results establish scaling principles for reranking and provide actionable insights for building industrial-grade retrieval systems.

TARSE: Test-Time Adaptation via Retrieval of Skills and Experience for Reasoning Agents

Complex clinical decision making often fails not because a model lacks facts, but because it cannot reliably select and apply the right procedural knowledge and the right prior example at the right reasoning step. We frame clinical question answering as an agent problem with two explicit, retrievable resources: skills, reusable clinical procedures such as guidelines, protocols, and pharmacologic mechanisms; and experience, verified reasoning trajectories from previously solved cases (e.g., chain-of-thought solutions and their step-level decompositions). At test time, the agent retrieves both relevant skills and experiences from curated libraries and performs lightweight test-time adaptation to align the language model's intermediate reasoning with clinically valid logic. Concretely, we build (i) a skills library from guideline-style documents organized as executable decision rules, (ii) an experience library of exemplar clinical reasoning chains indexed by step-level transitions, and (iii) a step-aware retriever that selects the most useful skill and experience items for the current case. We then adapt the model on the retrieved items to reduce instance-step misalignment and to prevent reasoning from drifting toward unsupported shortcuts. Experiments on medical question-answering benchmarks show consistent gains over strong medical RAG baselines and prompting-only reasoning methods. Our results suggest that explicitly separating and retrieving clinical skills and experience, and then aligning the model at test time, is a practical approach to more reliable medical agents.

Learning to Reason for Multi-Step Retrieval of Personal Context in Personalized Question Answering

Personalization in Question Answering (QA) requires answers that are both accurate and aligned with users' background, preferences, and historical context. Existing state-of-the-art methods primarily rely on retrieval-augmented generation (RAG) solutions that construct personal context by retrieving relevant items from the user's profile. Existing methods use the user's query directly to retrieve personal documents, and such strategies often lead to surface-level personalization. We propose PR2 (Personalized Retrieval-Augmented Reasoning), a reinforcement learning framework that integrates reasoning and retrieval from personal context for personalization. PR2 learns adaptive retrieval-reasoning policies, determining when to retrieve, what evidence to retrieve from user profiles, and how to incorporate it into intermediate reasoning steps. By optimizing multi-turn reasoning trajectories under a personalized reward function, the framework reinforces reasoning paths that better align with user-specific preferences and contextual signals reflected by the reward model. Extensive experiments on the LaMP-QA benchmark using three LLMs show that PR2 consistently outperforms strong baselines, achieving an average relative improvement of 8.8%-12% in personalized QA.

Scaling Laws for Embedding Dimension in Information Retrieval

Dense retrieval, which encodes queries and documents into a single dense vector, has become the dominant neural retrieval approach due to its simplicity and compatibility with fast approximate nearest neighbor algorithms. As the tasks dense retrieval performs grow in complexity, the fundamental limitations of the underlying data structure and similarity metric -- namely vectors and inner-products -- become more apparent. Prior recent work has shown theoretical limitations inherent to single vectors and inner-products that are generally tied to the embedding dimension. Given the importance of embedding dimension for retrieval capacity, understanding how dense retrieval performance changes as embedding dimension is scaled is fundamental to building next generation retrieval models that balance effectiveness and efficiency. In this work, we conduct a comprehensive analysis of the relationship between embedding dimension and retrieval performance. Our experiments include two model families and a range of model sizes from each to construct a detailed picture of embedding scaling behavior. We find that the scaling behavior fits a power law, allowing us to derive scaling laws for performance given only embedding dimension, as well as a joint law accounting for embedding dimension and model size. Our analysis shows that for evaluation tasks aligned with the training task, performance continues to improve as embedding size increases, though with diminishing returns. For evaluation data that is less aligned with the training task, we find that performance is less predictable, with performance degrading with larger embedding dimensions for certain tasks. We hope our work provides additional insight into the limitations of embeddings and their behavior as well as offers a practical guide for selecting model and embedding dimension to achieve optimal performance with reduced storage and compute costs.

Inferential Question Answering

Despite extensive research on a wide range of question answering (QA) systems, most existing work focuses on answer containment-i.e., assuming that answers can be directly extracted and/or generated from documents in the corpus. However, some questions require inference, i.e., deriving answers that are not explicitly stated but can be inferred from the available information. We introduce Inferential QA -- a new task that challenges models to infer answers from answer-supporting passages which provide only clues. To study this problem, we construct QUIT (QUestions requiring Inference from Texts) dataset, comprising 7,401 questions and 2.4M passages built from high-convergence human- and machine-authored hints, labeled across three relevance levels using LLM-based answerability and human verification. Through comprehensive evaluation of retrievers, rerankers, and LLM-based readers, we show that methods effective on traditional QA tasks struggle in inferential QA: retrievers underperform, rerankers offer limited gains, and fine-tuning provides inconsistent improvements. Even reasoning-oriented LLMs fail to outperform smaller general-purpose models. These findings reveal that current QA pipelines are not yet ready for inference-based reasoning. Inferential QA thus establishes a new class of QA tasks that move towards understanding and reasoning from indirect textual evidence.

Improving User Privacy in Personalized Generation: Client-Side Retrieval-Augmented Modification of Server-Side Generated Speculations

Personalization is crucial for aligning Large Language Model (LLM) outputs with individual user preferences and background knowledge. State-of-the-art solutions are based on retrieval augmentation, where relevant context from a user profile is retrieved for LLM consumption. These methods deal with a trade-off between exposing retrieved private data to cloud providers and relying on less capable local models. We introduce $P^3$, an interactive framework for high-quality personalization without revealing private profiles to server-side LLMs. In $P^3$, a large server-side model generates a sequence of $k$ draft tokens based solely on the user query, while a small client-side model, with retrieval access to the user's private profile, evaluates and modifies these drafts to better reflect user preferences. This process repeats until an end token is generated. Experiments on LaMP-QA, a recent benchmark consisting of three personalized question answering datasets, show that $P^3$ consistently outperforms both non-personalized server-side and personalized client-side baselines, achieving statistically significant improvements of $7.4%$ to $9%$ on average. Importantly, $P^3$ recovers $90.3%$ to $95.7%$ of the utility of a ``leaky'' upper-bound scenario in which the full profile is exposed to the large server-side model. Privacy analyses, including linkability and attribute inference attacks, indicate that $P^3$ preserves the privacy of a non-personalized server-side model, introducing only marginal additional leakage ($1.5%$--$3.5%$) compared to submitting a query without any personal context. Additionally, the framework is efficient for edge deployment, with the client-side model generating only $9.2%$ of the total tokens. These results demonstrate that $P^3$ provides a practical, effective solution for personalized generation with improved privacy.

Can Instructed Retrieval Models Really Support Exploration?

Exploratory searches are characterized by under-specified goals and evolving query intents. In such scenarios, retrieval models that can capture user-specified nuances in query intent and adapt results accordingly are desirable -- instruction-following retrieval models promise such a capability. In this work, we evaluate instructed retrievers for the prevalent yet under-explored application of aspect-conditional seed-guided exploration using an expert-annotated test collection. We evaluate both recent LLMs fine-tuned for instructed retrieval and general-purpose LLMs prompted for ranking with the highly performant Pairwise Ranking Prompting. We find that the best instructed retrievers improve on ranking relevance compared to instruction-agnostic approaches. However, we also find that instruction following performance, crucial to the user experience of interacting with models, does not mirror ranking relevance improvements and displays insensitivity or counter-intuitive behavior to instructions. Our results indicate that while users may benefit from using current instructed retrievers over instruction-agnostic models, they may not benefit from using them for long-running exploratory sessions requiring greater sensitivity to instructions.

Beyond a Million Tokens: Benchmarking and Enhancing Long-Term Memory in LLMs

Evaluating the abilities of large language models (LLMs) for tasks that require long-term memory and thus long-context reasoning, for example in conversational settings, is hampered by the existing benchmarks, which often lack narrative coherence, cover narrow domains, and only test simple recall-oriented tasks. This paper introduces a comprehensive solution to these challenges. First, we present a novel framework for automatically generating long (up to 10M tokens), coherent, and topically diverse conversations, accompanied by probing questions targeting a wide range of memory abilities. From this, we construct BEAM, a new benchmark comprising 100 conversations and 2,000 validated questions. Second, to enhance model performance, we propose LIGHT-a framework inspired by human cognition that equips LLMs with three complementary memory systems: a long-term episodic memory, a short-term working memory, and a scratchpad for accumulating salient facts. Our experiments on BEAM reveal that even LLMs with 1M token context windows (with and without retrieval-augmentation) struggle as dialogues lengthen. In contrast, LIGHT consistently improves performance across various models, achieving an average improvement of 3.5%-12.69% over the strongest baselines, depending on the backbone LLM. An ablation study further confirms the contribution of each memory component.

Benchmarking Information Retrieval Models on Complex Retrieval Tasks

Large language models (LLMs) are incredible and versatile tools for text-based tasks that have enabled countless, previously unimaginable, applications. Retrieval models, in contrast, have not yet seen such capable general-purpose models emerge. To achieve this goal, retrieval models must be able to perform complex retrieval tasks, where queries contain multiple parts, constraints, or requirements in natural language. These tasks represent a natural progression from the simple, single-aspect queries that are used in the vast majority of existing, commonly used evaluation sets. Complex queries naturally arise as people expect search systems to handle more specific and often ambitious information requests, as is demonstrated by how people use LLM-based information systems. Despite the growing desire for retrieval models to expand their capabilities in complex retrieval tasks, there exist limited resources to assess the ability of retrieval models on a comprehensive set of diverse complex tasks. The few resources that do exist feature a limited scope and often lack realistic settings making it hard to know the true capabilities of retrieval models on complex real-world retrieval tasks. To address this shortcoming and spur innovation in next-generation retrieval models, we construct a diverse and realistic set of complex retrieval tasks and benchmark a representative set of state-of-the-art retrieval models. Additionally, we explore the impact of LLM-based query expansion and rewriting on retrieval quality. Our results show that even the best models struggle to produce high-quality retrieval results with the highest average nDCG@10 of only 0.346 and R@100 of only 0.587 across all tasks. Although LLM augmentation can help weaker models, the strongest model has decreased performance across all metrics with all rewriting techniques.