The generation quality of large language models (LLMs) is often improved by utilizing inference-time scaling methods such as Chain-of-Thought (sequential scaling) or Self-Consistency (parallel scaling). We introduce Roster of Experts (RoE), a training-free inference algorithm for Mixture-of-Experts (MoE) models that defines a third, orthogonal scaling direction. Rather than generating longer outputs or aggregating multiple sequences, RoE improves the next token prediction distribution by treating a single MoE as a dynamic ensemble of internal subnetworks. It injects controlled stochasticity into the expert routing mechanism via Gumbel-Top-k sampling, enabling the model to aggregate multiple diverse subnetworks into a single, higher-quality token prediction. Crucially, because RoE improves the model’s output distribution rather than sampling from it, it is the only approach in this family that is applicable to open-ended generation tasks where sequence-level aggregation is infeasible. Furthermore, RoE is orthogonal to existing scaling methods; combining RoE with Self-Consistency yields greater accuracy than either method alone. To overcome the computational cost of multiple forward passes at inference, we employ an efficient batching strategy that exploits hardware parallelism to minimize latency, requiring no fine-tuning of model parameters.

Mapping local news coverage from textual content is a challenging problem that requires extracting precise location mentions from news articles. While traditional named entity taggers are able to extract geo-political entities and certain non geo-political entities, they cannot recognize precise location mentions such as addresses, streets and intersections that are required to accurately map the news article. We fine-tune a BERT-based language model for achieving high level of granularity in location extraction. We incorporate the model into an end-to-end tool that further geocodes the extracted locations for the broader objective of mapping news coverage.