small-1-7b-llm-fine-tuned-act-ensemble-router-assigns CLM-0639observationsingle paperReviewed 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.
Capability: Predicting future events
Observed on
Univariate forecasting on eight standard benchmarks (input and horizon 96), Qwen3-1.7B router, candidate pools of 4 small or foundation forecasters, oracle weights from a train-set.
Sources
- Validated by: Average MSE across eight benchmarks after SFT: Rule-CoT 0.1210 vs GPT-CoT 0.1260 on the small-model candidate group (better on 6/8 datasets) and 0.1455 vs 0.1456 on the foundation-model group; ablating CoT raised MSE from 0.1210 to 0.1315 in-distribution and from 0.1719 to 0.2675 out-of-distribution.. Measured on eight univariate benchmarks with one backbone and one pool size as default; [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 data-preparation pipelines in specialist domains (math, finance, law, medicine), diagnosing a small pilot set of generated samples against domain-selected quality criteria and rewriting the individual pipeline components (chunking, instruction builder, response generator) before full-scale generation yields training data with higher downstream fine-tuning utility than post-hoc filtering or one-shot agent/skill generators using the same backbone — but the gain is non-monotonic in rounds, peaking at two and degrading after.Improving its own scaffolding from failures · reviewed by AI
- When an agent builds a whole Python repository from natural-language requirements alone, refining component boundaries with explicit cohesion and coupling thresholds until no split or merge is triggered produces higher functionality coverage and test pass rate than letting the LLM decide restructuring actions for a fixed budget of rounds, because unconstrained LLM-decided actions keep over-decomposing past the point of a coherent modular structure.Generating and editing working code · 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
- For small open-weight backbones (Qwen3-4B) driving long-horizon remote-sensing tool chains, inserting a step-level verifier that checks each intermediate tool output against schema, sub-goal and domain constraints and triggers replanning is the single largest architectural gain, while hierarchical decomposition without domain-aligned tool grouping adds little and can lower final accuracy relative to a flat agent with the same verifier.Completing tasks that take many steps or many hours · reviewed by AI
- On repository-grounded LLM-infrastructure engineering tasks, frontier agents score far below ceiling (best 36.53% overall), and every one of eight models scores lower on long-horizon multi-file implementation than on single-file kernel completion, even though the full repository and test cases are visible.Completing tasks that take many steps or many hours · 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: low. Falsifier as drafted: Running the same pipeline with API-generated CoT and finding it beats rule-templated CoT on held-out datasets, or finding that removing CoT entirely costs no more out-of-distribution than in-distribution.