Notes to Self: Can LLMs Benefit from Experiential Abstractions?
Humans distill experience into reusable abstractions, e.g., strategies and cautionary reminders, and apply them to gradually solve problems more effectively. We study whether Large Language Models (LLMs) can similarly benefit from such experiential abstractions. From LLMs' solution traces on the MATH training set, a stronger teacher or the LLMs themselves extract natural-language abstractions into a retrievable library. We explore two usage modes: (1) inference-time retrieval and (2) reinforcement learning (RL) with abstraction-augmented training prompts. Experiential abstractions improve LLM
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- FuzzySimilar title/name (fuzzy) · 87%pbek/QOwnNotes →
“Fuzzy title match (0.94): “Notes to Self: Can LLMs Benefit from Experiential Abstractio” ≈ “pbek/QOwnNotes””
- FuzzyOverlapping authors or contributors · 62%modular/modular →
“Shared author/contributor keys: liu”
- LinkedLinked via arxiv author · 85%Chang Liu →
“Notes to Self: Can LLMs Benefit from Experiential Abstractions?”
- LinkedLinked via arxiv author · 85%Xinyu Li →
“Notes to Self: Can LLMs Benefit from Experiential Abstractions?”
- LinkedLinked via arxiv author · 85%Artur Dubrawski →
“Notes to Self: Can LLMs Benefit from Experiential Abstractions?”