#01Oct 21, 2026
cs.IR
pplx-embed-v2-late Technical Report
Markus Krimmel, Louis Milliken, Ekaterina Glazkova and 8 more
In this report, we introduce pplx-embed-v2-late, a pair of 0.6B and 9B multilingual and multimodal late-interaction models. Both models achieve state-of-the-art results within their respective size classes across a range of text, image, and visual document retrieval benchmarks. On visual document retrieval tasks, pplx-embed-v2-late-0.6b is competitive with models several times larger than itself. The models are trained via representation distillation from an 18B teacher. Consequently, they share an embedding space that supports asymmetric deployment: documents can be indexed with the 9B model and queried with the low-latency 0.6B model. Both models are available at https://huggingface.co/collections/perplexity-ai/pplx-embed-v2.
#02Oct 21, 2026
cs.CL
A Strong MoE Is All You Need for Process Supervision
Guanzhi Deng, Zengyan Liu, Haibo Wang and 5 more
Process reward models (PRMs) provide fine-grained supervision for multi-step reasoning, yet obtaining reliable process labels remains a central bottleneck. Existing approaches typically rely on human annotation, rollout-based estimation, or auxiliary verification. In this work, we explore a different source of process supervision natively exposed by Mixture-of-Experts (MoE) models: expert routing. We show that simple routing statistics provide an intrinsic process-confidence signal that is competitive with conventional output-logit confidence. Building on this observation, we construct routing confidence by combining overall routing mass, head routing mass, and within-step routing variation, and introduce RoCo-Rethink, which identifies the least confident reasoning step and selectively regenerates the subsequent trajectory. We further introduce RoCo-PRM, which uses a strong MoE to automatically annotate reasoning trajectories with routing-derived step scores and distills them into a standalone PRM. Across multiple MoE architectures and diverse reasoning tasks, both RoCo-Rethink and RoCo-PRM consistently outperform strong baselines, while RoCo-PRM transfers effectively across both model architectures and tasks. Our findings suggest that expert routing is more than a computational dispatch mechanism: it provides an architecture-native source of process supervision for reasoning.
#03Oct 21, 2026
cs.LG
Harness Scaling for Generalizable Agentic Reinforcement Learning
Fei Bai, Huatong Song, Chuan Hao and 5 more
Agent harnesses have become a central component of modern agentic systems, shaping how models use tools, manage context, and coordinate long-horizon interaction. Realizing the benefits of these harnesses depends on how effectively the underlying model can make use of them, motivating reinforcement learning directly through the harness. However, existing approaches typically train through a single fixed harness, which can lead to specialization and limited transfer across harness configurations. To address this, we introduce \textbf{Harness Scaling}, a new scaling dimension for agentic RL that extends training from a single harness to diverse execution configurations. Concretely, we construct a factorized harness space, select a compact training subset with balanced factorial coverage, and organize cross-harness training to combine global coverage with controlled variation for the same task. Across multiple benchmarks, model scales, and RL algorithms, Harness Scaling consistently improves generalization to unseen white-box configurations. Moreover, models trained through structured white-box variation transfer effectively to unseen external black-box harnesses, outperforming models trained through a single complex black-box harness. These results establish harness diversity as a meaningful scaling dimension for agentic RL and highlight the importance of how this diversity is organized for cross-harness generalization.
#04Oct 21, 2026
cs.LG
Stochastic Pruning of the DPO Computation Graph: 1.4x Faster Training and Lower Peak Memory
Yubing Wu, Wei Chen, Junmei Yang and 2 more
Direct preference optimization (DPO) incurs training costs in output scoring, sequence execution, and backpropagation, so reducing supervision alone does not necessarily save computation. We introduce a unified framework for stochastic pruning of the DPO computation graph that independently samples response positions for scoring, response prefixes for execution, and preference branches for backpropagation. By making these selections before the corresponding operations, the framework omits computation at each stage of a training step. Our analysis characterizes the objective changes induced by position and sequence pruning and uses inverse-probability weighting to obtain a conditionally unbiased single-branch gradient estimator for the sampled objective. Controlled measurements show that position and sequence pruning primarily reduce peak memory, with most output-memory savings attributable to an implementation that retains full response supervision. Branch pruning yields a $1.426\times$ step-time speedup over a concatenated DPO baseline and reduces training time by about 30\% relative to stock DPO across two model families, with slightly lower mean accuracy and benchmark-dependent changes in open-ended scores. These findings relate the resource benefits of stochastic graph pruning to the computation omitted and the resulting changes in the training objective.
#05Oct 21, 2026
cs.CL
OmniSimulator: Aligning Small Language Models for Authentic Heterogeneous Behavior Modeling
Jiawei Chen, Ruoxi Xu, Boxi Cao and 12 more
Large language models (LLMs) have shown great potential as general-purpose user simulators for interactive systems. Prior studies have demonstrated that training on high-quality data can substantially improve model reasoning and downstream task performance. However, the upper bound of user simulation is still constrained by the model's reasoning paradigm: native chain-of-thought often degenerates into shallow pattern matching and fails to adequately model the behavioral factors underlying real user decisions. In this paper, we propose OmniSimulator, a novel training framework designed to improve the predictive ability of small LLMs for realistic user simulation on short-video platform environments. Specifically, OmniSimulator organizes behavior-oriented reasoning through a structured schema with five dimensions, carefully selects data from real user-behavior scenarios, and introduces a teacher model to generate high-quality reasoning traces under the proposed schema, helping small models learn behavior-relevant reasoning patterns rather than relying on superficial behavioral correlations. Based on this process, we construct 9,881 high-quality training instances from three months of historical interaction trajectories of 200 real users, averaging about 50 distilled examples per user. Experimental results show that OmniSimulator improves the average score of Qwen3-8B by 34.33%, achieves performance competitive with Claude-Sonnet-4.5, and supports smaller-scale, self-hosted deployment. These results suggest that learning structured behavior-oriented reasoning patterns is a promising direction towards scalable and high-fidelity LLM-based user simulation.