{"paper":{"title":"LLM Flow Processes for Text-Conditioned Regression","license":"http://creativecommons.org/licenses/by/4.0/","headline":"LLM regression outputs become better calibrated and trajectory-consistent when blended with a lightweight diffusion neural process.","cross_cats":["cs.CL","stat.ML"],"primary_cat":"cs.LG","authors_text":"Felix Biggs, Samuel Willis","submitted_at":"2026-01-05T21:20:38Z","abstract_excerpt":"Recent work has demonstrated surprisingly good performance of pre-trained LLMs on regression tasks (for example, time-series prediction), with the ability to incorporate expert prior knowledge and the information contained in textual metadata. However we observe major error cascades even in short sequences < ~100 points; these models are also computationally intensive and difficult to parallelise. Marginal LLM predictions do not suffer this issue and are trivially parallelised, but can predict over-broad densities. To address this, we propose combining these densities with a lightweight (diffu"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Combining these densities with a lightweight (diffusion-based) neural process leads to better-calibrated predictions overall, outputs locally consistent trajectories, and leads to text-conditioned function space selection in the meta-learner. We propose a gradient-free (and non-Monte Carlo) method for sampling from a product-of-experts of a score model and an 'expert' (here the LLM predictive densities).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that marginal LLM predictions can be convolved with a Gaussian in closed form, enabling the proposed gradient-free sampling method, and that the lightweight neural process can correct over-broad densities without introducing new inconsistencies or requiring extensive tuning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LLM densities fused with a diffusion neural process yield better-calibrated, locally consistent text-conditioned regression predictions plus a gradient-free product-of-experts sampler.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"LLM regression outputs become better calibrated and trajectory-consistent when blended with a lightweight diffusion neural process.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"79ba12074260a7eb5ce4d432bd19735a79bcbfffbbc3fb3cdefe6cffbc16057a"},"source":{"id":"2601.06147","kind":"arxiv","version":2},"verdict":{"id":"5f768847-73d6-40ef-a0c9-7eefc01e51b5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T17:14:53.058736Z","strongest_claim":"Combining these densities with a lightweight (diffusion-based) neural process leads to better-calibrated predictions overall, outputs locally consistent trajectories, and leads to text-conditioned function space selection in the meta-learner. We propose a gradient-free (and non-Monte Carlo) method for sampling from a product-of-experts of a score model and an 'expert' (here the LLM predictive densities).","one_line_summary":"LLM densities fused with a diffusion neural process yield better-calibrated, locally consistent text-conditioned regression predictions plus a gradient-free product-of-experts sampler.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that marginal LLM predictions can be convolved with a Gaussian in closed form, enabling the proposed gradient-free sampling method, and that the lightweight neural process can correct over-broad densities without introducing new inconsistencies or requiring extensive tuning.","pith_extraction_headline":"LLM regression outputs become better calibrated and trajectory-consistent when blended with a lightweight diffusion neural process."},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ceae6dcd7ae15cbff9713562413a110ffcb3c4744957a497f9ac1421c952385d"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}