frozen-pretrained-time-series-forecaster-leaves-structured-residual-error-having CLM-0573mechanismsingle paperReviewed by AI
When a frozen pretrained time-series forecaster leaves structured residual error, having an LLM propose named semantic features of that residual (e.g. renewable generation over total load) and admitting them through a validation gate into a small post-hoc corrector reduces test error more than statistical feature banks or compositional search over the raw channels, but the gain scales with how much raw error the backbone leaves and vanishes once the backbone already fits the series.
Capability: Predicting future events
Observed on
Six public forecasting datasets, six frozen pretrained backbones, covariates routed only to the corrector; holds under two LLM backends and on top of fine-tuned backbones; no relia.
Sources
- supportsWhen Do Corrective Features Help? An Agent for Corrective Feature Discovery on Black-Box ForecastersValidated by: Mean reduction in test wMAPE versus the raw frozen backbone across 36 backbone-dataset cells under a rolling-origin, multi-seed protocol, against three external feature-engineering systems (TSFresh, CAAFE, LLM-FE) run through the identical gate and corrector, plus a no-LLM ablation, a covariate-only corrector, and a head-only fine-tune; paired Wilcoxon p<0.01, up to 27% error reduction on the weakest backbones, and the advantage holds at every feature budget K.. [truncated]
Either way, bring evidence: a link, or what you saw.
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- For LLM-based forecasting systems, reported benchmark gains often cannot be attributed to temporal reasoning, because test periods predate the training cutoff, ablations show the LLM component sometimes adds nothing over the statistical or retrieval parts, and accuracy shifts under small input perturbations.Predicting future events · reviewed by AI
- When a small (1.7B) LLM is fine-tuned to act as an ensemble router that assigns per-sample weights across candidate time-series forecasters, chain-of-thought supervision built by rule-templating known oracle weights against extracted temporal features matches or beats CoT generated by a frontier API model, and removing the CoT step costs far more on unseen candidate model pools than on in-distribution ones.Predicting future events · reviewed by AI
- Across LLM-based forecasting systems, reported accuracy gains frequently cannot be attributed to the language model's temporal reasoning, because evaluation windows predate training cutoffs, component ablations show the LLM sometimes adds nothing over the statistical or retrieval parts, and results shift under small input perturbations.Predicting future events · reviewed by AI
- When a text-pretrained LLM is prompted to forecast a numeric price series, rounding the input values to integers while keeping the original price scale gives lower forecast error than full-precision raw values, training-set standardization, or per-window instance normalization, because normalized values fall outside the numeric distributions the model saw in pretraining and per-window rescaling makes the same price appear differently across samples.Predicting future events · reviewed by AI
- When research-idea forecasts are scored by whether a later paper matches them under an LLM judge rubric, higher scores partly reflect broader, less specific ideas: the backbone with the higher hit rate also produced measurably more general forecasts and had more retrieved candidate papers pass the gate, and tightening the specificity threshold does not separate the two.Whether the measurement made the finding · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: A run on the same grid where the LLM-named features add no lift beyond compositional search and statistical banks at matched budget, or where lift is uncorrelated with the backbone's raw error.