Skip to content
AI News HubLIVE
Original source2 min read

Prompt-Space Meta-Learning Does Not Transfer Across Users: A Frozen-LLM Negative Result

Summary

An arXiv study uses Muse, a reflective prompt-evolution method that learns a shared adaptation prompt, to test whether prompt-space meta-learning encodes transferable cross-user personalization for frozen LLMs. Across 200 held-out users on LaMP-2 and LaMP-3, Muse does not significantly beat its un-evolved seed prompt or a structure-broken control, and is dominated by plain few-shot retrieval on rating (ΔMAE +0.175). The authors attribute this to meta-objective collapse.

SourcearXiv Machine LearningAuthor: Liam Byrne, David Dylan, Orla Fitzgerald, Eoin Doyle, Ciara Nolan, Padraig Lynch, Sinead Gallagher
Prompt-Space Meta-Learning Does Not Transfer Across Users: A Frozen-LLM Negative Result
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 25 Jun 2026]

Title:Prompt-Space Meta-Learning Does Not Transfer Across Users: A Frozen-LLM Negative Result

View a PDF of the paper titled Prompt-Space Meta-Learning Does Not Transfer Across Users: A Frozen-LLM Negative Result, by Liam Byrne and 6 other authors

View PDF HTML (experimental)

Abstract:Personalizing a frozen large language model (LLM) to individual users is often framed as a meta-learning problem in prompt space: each user is a task, and one seeks a shared natural-language adaptation policy that, given a handful of the user's labeled interactions, configures the frozen model for that user. The framing is attractive because it is backbone-agnostic and reuses the machinery of prompt optimization, yet the field rarely tests whether the optimized meta-objective encodes transferable cross-user adaptation rather than generic instruction quality. We study this question with Muse (Meta-learned User-adaptation via Shared Evolution), which evolves a single shared adaptation prompt over a meta-train user population by reflective prompt evolution, freezes it, and applies it zero-shot to held-out users; matched controls isolate learning from confounds of phrasing and selection. On two standard personalization benchmarks (LaMP-2 categorization and LaMP-3 rating) over 200 held-out users each, Muse does not significantly improve on its own un-evolved seed prompt or on a structure-broken control that meta-trains on mismatched user-support pairs, and is dominated by plain few-shot retrieval on the rating task (Delta MAE +0.175, p

new | recent | 2026-09

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

IArxiv recommender toggle

IArxiv Recommender (What is IArxiv?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Muse evolves one shared adaptation prompt over a meta-train user population, freezes it, and applies it zero-shot to held-out users
  • On 200 held-out users per benchmark, Muse does not significantly improve over its seed prompt or a mismatched user-support control
  • Plain few-shot retrieval significantly beats Muse on LaMP-3 rating (ΔMAE +0.175, p < 0.001)
  • The meta-validation objective is statistically invariant to genuine user-support correspondence (p=0.555 and p=0.622), causing meta-objective collapse

Highlights and analysis are generated automatically and may contain errors. Check the original source.