Research map / Large language models / Instruction and fine-tuning
Language Models are Weak Learners
Hariharan Manikandan, Yiding Jiang, J Zico Kolter
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use summary-based weak learners with AdaBoost instead of a single Summary predictor | GPT-3 Curie tabular classification · 18 UCI and OpenML datasets | test error | Improved |
| Replace Few-shot examples with a generated Summary | GPT-3 Curie tabular classification · Most tested datasets | test error | Improved |
| Use five bins with low/medium/high quantifiers for continuous attributes (ablation) | Summary · Tabular datasets | summary quality | Improved |
| Use actual attribute names rather than dummy feature names (ablation) | Summary · Tabular classification | few-shot performance | Improved |
| Increase support examples in the prompt (ablation) | Few-shot · Tabular classification | test performance | Mixed |
| Increase support examples summarized into the context (ablation) | Summary · Tabular classification | test performance | Improved |
| Use GPT-generated rather than templatized data descriptions (ablation) | Summary · Tabular classification | test performance | Improved |
| Use cluster sampling rather than random sampling (ablation) | Summary Boosting · Tabular classification | rounds to convergence | Improved |
| Use cluster rather than random sampling after sufficient boosting rounds (ablation) | Summary Boosting · Tabular classification | final test error | No clear effect |
| Present examples grouped by class rather than shuffled (ablation) | Summary · Four datasets | test performance | No clear effect |
| Use explicit summary instructions rather than a tl;dr prompt (ablation) | Summary · Tabular classification | test performance | No clear effect |
| Use chain-of-thought inference rather than a prefix prompt (ablation) | Summary · Tabular classification | test performance | No clear effect |
Closest papers by meaning
- Escaping Collapse: The Strength of Weak Data for Large Language Model Training (NeurIPS 2025 · similarity 0.91)
- Large Language Models Can Automatically Engineer Features for Few-Shot Tabular Learning (ICML 2024 · similarity 0.88)
- Large Language Models are Good Prompt Learners for Low-Shot Image Classification (CVPR 2024 · similarity 0.88)
- PREFER: Prompt Ensemble Learning via Feedback-Reflect-Refine (AAAI 2024 · similarity 0.88)
- Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights (NeurIPS 2025 · similarity 0.88)
- Weak-to-Strong Search: Align Large Language Models via Searching over Small Language Models (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
