#01Nov 6, 2026
cs.AI
Knowing When to Intervene: Causal Evaluation of AI Agents
Ayush Sawarni, Wenqian Xing, Tijana Zrnic and 2 more
Current evaluations of agentic AI systems primarily rely on benchmarks. However, the static nature of benchmarks makes them uninformative about the value of auditing interventions at deployment time. For reliable deployment, we often need to determine within an agent's trajectory whether or not to intervene, e.g., by verifying the agent's current output or steering it toward different behavior, using either human review or more costly models or time-consuming agentic supervision. The value of such interventions is fundamentally a causal inference question, asking whether the causal effect of such an intervention is beneficial for the final outcome. Estimating this causal effect is particularly challenging due to the dynamic nature of the problem: the impact of an intervention depends on the current context and environment, as well as the agent's future behavior. We develop the first methodologically grounded framework for evaluating the causal effects of randomized interventions within agent trajectories. By leveraging a relatively small dataset of agent trajectories collected under randomized interventions such as verification or screening, we can estimate the dynamic causal effect of a rich class of (contextual) intervention policies. The structural parameters that we estimate can also be used to construct safety leaderboards that capture the risk of various agentic actions, which can be used by practitioners for audit prioritization. Applying our methods to tau-bench Retail, ToolSandbox, and ImpossibleBench, we find that intervention effects vary substantially across agent actions and states. More broadly, our work suggests a rich research agenda at the intersection of causal inference and agentic evaluation, which we believe is essential for the reliable and safe deployment of AI agents in the real world.
#02Nov 6, 2026
cs.AI
CPU-ARC: Agentic Reasoning Caching for on-CPU Reasoning Reuse
Peter Belcak, Greg Heinrich, Saurav Muralidharan and 2 more
Large language models reinvent the wheel all the time. They conflate task-solving with the execution of the solutions, which forces them to solve every problem from scratch. We show that once a task has been reasoned through, the application of its solution to a new problem instance can be done very efficiently and with little involvement from language models. We introduce agentic reasoning caching, a method for constructing robust CPU-executable agentic reasoning caches (ARCs) from LLM reasoning traces. ARCs show on-par performance with frontier LLMs across a range of difficult mathematical and scientific tasks, are fully explainable, continually trainable, robust to input perturbations, and 16-291x cheaper to invoke than the source model.
#03Nov 6, 2026
cs.CL
ROSE: Risk-Aware Orthogonal Subspace Navigation for Lifelong Knowledge Editing in Multimodal Large Language Models
Lingyun Song, Ziyao Chen, Kang Pan and 5 more
Multimodal Large Language Models (MLLMs) require continuous factual updates to stay current, yet lifelong knowledge editing remains a daunting challenge. The primary obstacle is the severe Locality Erosion and Catastrophic Forgetting, where sequential edits inevitably interfere with pre-trained capabilities and prior updates. In this paper, we propose \textbf{ROSE}, a \textbf{R}isk-aware \textbf{O}rthogonal \textbf{S}ubspace navigation framework for lifelong knowledge \textbf{E}diting. ROSE addresses forgetting by aiming to restrict parameter updates to the orthogonal complement of previously edited subspaces. Crucially, it introduces a risk-guided mechanism that utilizes a dynamically computed risk mask to safeguard specific factual parameter subspaces essential for model stability. During inference, ROSE employs a multi-granularity knowledge integration strategy, featuring prototype-anchored semantic gating and perturbation-driven subspace fusion to precisely activate edited knowledge for relevant queries while limiting interference on unrelated inputs. Comprehensive experiments on two prominent multimodal benchmarks show that ROSE is competitive with state-of-the-art approaches across five key metrics, while maintaining outstanding stability over $1,000$ sequential updates. Our code is available at https://github.com/sunlitsong/ROSE.
#04Nov 6, 2026
cs.AI
On the Reliability of Majority-Vote Pseudo-Labels in Test-Time Reinforcement Learning
Haoyu Wang, Ru Wang, Yuanyuan Wang and 2 more
Test-time reinforcement learning (TTRL) adapts pretrained language models to unlabeled target-domain data using self-generated pseudo-labels, whose reliability remains poorly understood. We analyze majority-vote supervision for TTRL with online Direct Preference Optimization (DPO) and decompose its error into population-majority bias and prompt-dependent finite-rollout variation. Because population-majority bias is generally unidentifiable without external supervision, we focus on mitigating the downstream effects of finite-rollout variation. We propose \emph{Prompt-Conditional Distributionally Robust DPO (PC-Dr.DPO)}, which conditions within-prompt robust reweighting on observed vote reliability to make updates from sampling-sensitive prompts more conservative. We further introduce a cache-based mechanism to reduce self-supervised forgetting during online adaptation. Across six mathematical and knowledge-intensive reasoning benchmarks and five model backbones, PC-Dr.DPO matches or outperforms the strongest TTRL baseline in 24 of 30 settings.
#05Nov 6, 2026
cs.AI
Generative Latent Priors for Interpreting and Steering Protein Language Models
Daniel Hinjos, Lucía Urcelay
Dominant methods for interpreting large models impose strong structural assumptions. Generative Latent Priors (GLP) are a complementary approach: a diffusion model that learns the distribution of a frozen model's activations directly, imposing no assumptions. We port GLP to ESM-2 activations (UniRef50) and evaluate it at two host scales (ESM-2-8M and -650M), both as a lens for protein representations and as a prior for editing them. Generation quality scales smoothly with depth and compute. As a lens, single GLP "meta-neurons" are the strongest probes of discrete Swiss-Prot concepts at both hosts, beating InterPLM's released SAE features, and lead on continuous properties at 8M though not at 650M; falsification tests show this is mostly concentration, not new information. As an editing prior, the on-manifold correction makes DiffMean steering of three biochemical properties effective and restores on-manifoldness, but only in isolation: re-embedding the sequences actually produced after several rounds of decoding, the structural plausibility is compromised. This is mainly reproduced in additional SAE-based clamping structural steering experiments, where only Helix motif succeeds. The naturalness failure is traced to the rewriting coverage and the decoder rather than the prior, highlighting what the method must clear before it can serve for both discovery-by-editing and discovery-by-exploration. Trained GLPs are released on huggingface.co/collections/HPAI-BSC/glprot.