mlsanityData to 14 Oct 2026

The Prompt is Mightier than the Example

arXiv 2505.18485 ↗First posted 24 May 2025cs.LGcited by 0

On the map: LLM reasoning → Prompting and chain-of-thought reasoning

Abstract (arXiv)

Numerous recent prompt optimization approaches like chain-of-thought, have been demonstrated to significantly improve the quality of content generated by large language models (LLMs). In-context learning (ICL), a recent paradigm where a few representative examples guide content generation has also led to strong improvements in generation quality of LLM generated content. This idea has been applied to great effect in synthetic tabular data generation, where LLMs, through effective use of ICL and prompt optimization, can generate data that approximate samples from complex, heterogeneous distributions based on representative examples. However, ensuring high-fidelity synthetic data often requires a very large number of ICL examples which may be unavailable or costly to obtain. At the same time, as LLMs get larger and larger, their in-built prior knowledge becomes vast and can potentially substitute for specific data examples. In this paper, we introduce Knowledge-Guided Prompting (KGP) as a new knob in prompt optimization and explore the ability of KGP-based prompt optimization to offset the cost of ICL. Specifically, we explore the question `how many examples can a prompt substitute for?' and explore knowledge-guided prompting (KGP) where domain knowledge, either inferred or available, is explicitly injected into the prompt, reducing dependence on ICL examples. Our experiments systematically explore the trade-off between ICL and KGP, revealing an empirical scaling law that quantifies how quality of generated synthetic data varies with increasing domain knowledge and decreasing example count. Our results demonstrate that knowledge-guided prompting can be a scalable alternative, or addition, to in-context examples, unlocking new approaches to synthetic data generation.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning. Builds on 20 of its 47 references that are in the map. First posted on arXiv 24 May 2025.

Builds on

20 of 47 references.

References · 20

  1. 2022Generative Adversarial Networkscited by 5,627
  2. 2024RULER: What's the Real Context Size of Your Long-Context Language Models?cited by 1,346
  3. 2019Modeling Tabular data using Conditional GANcited by 699
  4. 2017Real-valued (Medical) Time Series Generation with Recurrent Conditional GANscited by 333
  5. 2022Language Models are Realistic Tabular Data Generatorscited by 294
  6. 2022Why Can GPT Learn In-Context? Language Models Implicitly Perform Gradient Descent as Meta-Optimizerscited by 204
  7. 2021How Many Data Points is a Prompt Worth?cited by 198
  8. 2017Generating Multi-label Discrete Patient Records using Generative Adversarial Networkscited by 197
  9. 2023Mixed-Type Tabular Data Synthesis with Score-based Diffusion in Latent Spacecited by 186
  10. 2018Data Synthesis based on Generative Adversarial Networkscited by 185
  11. 2022LIFT: Language-Interfaced Fine-Tuning for Non-Language Machine Learning Taskscited by 148
  12. 2023REaLTabFormer: Generating Realistic Relational and Tabular Data using Transformerscited by 94
8 more
  1. 2024Why Does the Effective Context Length of LLMs Fall Short?cited by 63
  2. 2023Generative Table Pre-training Empowers Models for Tabular Predictioncited by 55
  3. 2023TabuLa: Harnessing Language Models for Tabular Data Synthesiscited by 42
  4. 2021TabularNet: A Neural Network Architecture for Understanding Semantic Structures of Tabular Datacited by 36
  5. 2024TabEBM: A Tabular Data Augmentation Method with Distinct Class-Specific Energy-Based Modelscited by 16
  6. 2023Towards Understanding How Transformers Learn In-context Through a Representation Learning Lenscited by 15
  7. 2020STAN: Synthetic Network Traffic Generation with Generative Neural Modelscited by 11
  8. 2020Improving Movement Predictions of Traffic Actors in Bird's-Eye View Models using GANs and Differentiable Trajectory Rasterizationcited by 9

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024An Automatic Prompt Generation System for Tabular Data Tasks
  2. 2024Efficient Knowledge Injection in LLMs via Self-Distillation
  3. 2025SIPDO: Closed-Loop Prompt Optimization via Synthetic Data Feedback
  4. 2025Multi-Sample Prompting and Actor-Critic Prompt Optimization for Diverse Synthetic Data Generation
  5. 2024PromptKD: Distilling Student-Friendly Knowledge for Generative Language Models via Prompt Tuning
  6. 2025TRPrompt: Bootstrapping Query-Aware Prompt Optimization from Textual Rewards
  7. 2025The Art of Asking: Multilingual Prompt Optimization for Synthetic Data
  8. 2024Self-Instructed Derived Prompt Generation Meets In-Context Learning: Unlocking New Potential of Black-Box LLMs