Learning desk

MultiAgent EDU StackGather good sources. Teach what matters.
T5TauricResearch/TradingAgentsT5A Man Who Invented Modern AI (Before Everyone Else) – Jürgen Schmidhuber [video]T5GPT-4 finished training four years ago todayT5AI Settles a 25 Year-Old Problem We Left BehindT5What it was like working on LLMs and security at Meta (2022-2026)T5Ask HN: How do you go from writing code to deploying with agents?T5What Happened: OpenAI and HuggingFaceT5Apple says Mac users in China can connect to Alibaba's Qwen AI serviceT5Show HN: Try Benzi – A coding harness/agent beating Claude Code itself on SonnetT5The AI Apocalypse Is HereT3Auto mode is now the default in Claude Code for Pro, Max, and Team plansT5Show HN: Tura – Build agent that uses 80% less token and delivers better resultsT5TauricResearch/TradingAgentsT5A Man Who Invented Modern AI (Before Everyone Else) – Jürgen Schmidhuber [video]T5GPT-4 finished training four years ago todayT5AI Settles a 25 Year-Old Problem We Left BehindT5What it was like working on LLMs and security at Meta (2022-2026)T5Ask HN: How do you go from writing code to deploying with agents?T5What Happened: OpenAI and HuggingFaceT5Apple says Mac users in China can connect to Alibaba's Qwen AI serviceT5Show HN: Try Benzi – A coding harness/agent beating Claude Code itself on SonnetT5The AI Apocalypse Is HereT3Auto mode is now the default in Claude Code for Pro, Max, and Team plansT5Show HN: Tura – Build agent that uses 80% less token and delivers better results
← Dispatches

The Few-shot Dilemma: Over-prompting Large Language Models

Primary research

#872

T1new
Topic
unassigned (set during synthesis)
First seen
2026-07-28 07:16:45
Last seen
2026-07-28 07:16:45

Source raw items (1)

  • arXiv2026-07-28 07:15:59
    The Few-shot Dilemma: Over-prompting Large Language Models

    Over-prompting, a phenomenon where excessive examples in prompts lead to diminished performance in Large Language Models (LLMs), challenges the conventional wisdom about in-context few-shot learning. To investigate this few-shot dilemma, we outline a prompting framework that leverages three standard few-shot selection methods - random sampling, semantic embedding, and TF-IDF vectors - and evaluate these methods across multiple LLMs, including GPT-4o, GPT-3.5-turbo, DeepSeek-V3, Gemma-3, LLaMA-3.1, LLaMA-3.2, and Mistral. Our experimental results reveal that incorporating excessive domain-specific examples into prompts can paradoxically degrade performance in certain LLMs, which contradicts the prior empirical conclusion that more relevant few-shot examples universally benefit LLMs. Given the trend of LLM-assisted software engineering and requirement analysis, we experiment with two real-world software requirement classification datasets. By gradually increasing the number of TF-IDF-selected and stratified few-shot examples, we identify their optimal quantity for each LLM. This combined approach achieves superior performance with fewer examples, avoiding the over-prompting problem, thus surpassing the state-of-the-art by 1% in classifying functional and non-functional requirements.