text-pretrained-llm-prompted-forecast-numeric-price-series-rounding-input CLM-0248mechanismsingle paperReviewed 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.
Capability: Predicting future events
Observed on
Zero-shot prompting of a 4-bit quantized LLaMA-3 8B on hourly Bitcoin closing prices, 96-step input, 8-step horizon; one asset, one period (2023-2024)..
Sources
- Validated by: Validation-set MAPE/SMAPE/NRMSE on BTC/USDT hourly forecasting under a fixed zero-shot prompt with only the representation varied: rounded 3.01/3.05/0.05 vs raw 4.23/5.42/0.12, standard normalization 9.43/9.39/0.16, instance normalization 548.36/132.19/1.41; valid-output count unchanged at 1441.. Measured as a one-factor-at-a-time ablation with a fixed prompt, so representation is isolated, but on a single asset, single window length, single backbone, and a single 10% chronological validation split; [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
- When a text LLM is prompted to forecast numeric price series, feeding the history as integer-rounded values on the original price scale gives lower forecast error than full-decimal raw values, standard-normalized values, or per-window instance-normalized values, because rounding cuts token fragmentation while keeping magnitudes in the distribution the model saw in pretraining.Predicting future events · reviewed 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.Predicting future events · reviewed by AI
- On numeric sequences whose first differences repeat only after a long prefix, LLaMA 3.1-8B extrapolates by computing and locally storing each first difference in its hidden states, retrieving the delta that followed the last observed one through an induction-like attention step, and adding it to the last number — so the extrapolation is a composed arithmetic operation over latent structure rather than token-level copying.Digit-level arithmetic · reviewed by AI
- When an LLM judge is given a prior evaluation score as context metadata on a revision, its rubric scores shift toward that uninformative prior — seven of eight tested models (API and open-weight, 0-5 scale, 20 fixed texts) show a negative total shift, and on a categorical compliance task the metadata blocks 48% of error corrections and flips about 10% of previously correct judgments to the anchored wrong label.Judging another's work as well as a careful reviewer would · reviewed by AI
- 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
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: On other numeric series and other LLM backbones, instance- or standard-normalized inputs match or beat integer-rounded inputs under an otherwise identical prompt and decoding setup.