Are LLM emergent abilities real or measurement artifacts?
Do large language models develop sudden new capabilities at certain scales, or do discontinuous metrics just make gradual improvements look sudden? This matters because it changes how we predict and interpret model behavior.
The sharp, unpredictable transitions that define "emergent abilities" — capabilities appearing suddenly at certain model scales — are artifacts of the researcher's choice of metric rather than fundamental changes in model behavior.
The argument: nonlinear or discontinuous metrics (like exact string match) produce apparent emergent abilities, while linear or continuous metrics (like token edit distance) applied to the same model outputs show smooth, continuous, predictable changes with scale. The "emergence" lives in the measurement, not the model.
Three complementary validations:
- InstructGPT/GPT-3 family — tasks with claimed emergent abilities show smooth improvement under continuous metrics
- BIG-Bench meta-analysis — claimed emergent abilities evaporate with different metrics or better statistics
- Vision tasks — the same metric manipulation produces never-before-seen "emergent abilities" across diverse deep networks, confirming the mechanism is metric-dependent not domain-specific
This doesn't mean models don't improve with scale — they do, continuously. What it challenges is the narrative of sudden capability transitions that implies qualitative changes in what models can do. The practical implication: scaling predictions become much more tractable if improvements are smooth rather than discontinuous.
This connects to Do foundation models learn world models or task-specific shortcuts? — both challenge the narrative of fundamental capability leaps. Heuristics improve gradually with more data; emergence would require qualitative shifts. The metric artifact finding supports the heuristics interpretation.
Inquiring lines that read this note 36
This note is a source for these research framings, grouped by the broader line of inquiry each explores. Scan the bold lines of inquiry; follow any specific question forward.
Can language models reason beyond surface pattern matching? Do language models encode knowledge that influences generation, or primarily imitate surface patterns? What gaps exist between benchmark performance and real deployment outcomes?- What measurement artifacts emerge when annotators interpret the same question differently?
- What capability dimension does a closed-ended exam actually fail to measure?
- What causes models to develop domain capability cliffs after specialization?
- Why do metric choices constrain which model capabilities get developed?
- Do larger models develop more abstract features than smaller ones?
- What capabilities actually require massive scale versus specialized training regimes?
- Why do scaling laws show capability saturation at specific thresholds?
- Why does exemplar performance vary across order complexity diversity and style?
- Do rare cultural concepts fail predictably as model scale increases?
- How much do metric choices inflate claims about model capabilities?
- What language capabilities does fluency on standard benchmarks actually measure?
- Why do leaderboard metrics fail to capture human flourishing in LLM evaluation?
- How does measurement error in capability benchmarks systematically underestimate or overestimate true ability?
- Can a single competence score capture multiple separable dimensions of capability?
- How often do metric improvements fail to reflect real capability gains?
- Do emergent abilities result from genuine new capabilities or implicit in-context learning?
- How does trajectory burstiness compare to other structural properties that shape emergent capabilities?
- What makes some model capabilities reliable while others remain brittle?
- What capability boundary exists in LLM prediction of effect sizes?
Related concepts in this collection 3
This note in its neighbourhood — explore the map, then jump to a related concept in the list below.
Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph
-
Do foundation models learn world models or task-specific shortcuts?
When transformer models predict sequences accurately, are they building genuine world models that capture underlying physics and logic? Or are they exploiting narrow patterns that fail under distribution shift?
gradual heuristic improvement vs sudden capability emergence; both support the same underlying picture
-
Can neural networks learn compositional skills without symbolic mechanisms?
Do neural networks need explicit symbolic architecture to compose learned concepts, or can scaling alone enable compositional generalization? This asks whether compositionality is an architectural feature or an emergent property of scale.
appears to conflict: compositional generalization does emerge at scale, but may do so smoothly rather than suddenly
-
How much of LLM few-shot ability comes from training data?
Do large language models genuinely learn from a few examples, or are they mostly recognizing patterns from their training data? This matters for understanding what LLMs can actually do.
triple challenge to capabilities narrative: metric artifacts inflate emergence claims + task contamination inflates baselines + prompting techniques don't replicate
Related papers in this collection 8
Papers most semantically related to this note, ranked by cosine similarity in the embedding space.
- Are Emergent Abilities of Large Language Models a Mirage?
- Are Emergent Abilities in Large Language Models just In-Context Learning?
- Progress Measures For Grokking Via Mechanistic Interpretability
- MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselves
- Nested Learning: The Illusion of Deep Learning Architectures
- Interactive Evaluation Requires a Design Science
- xDailyBench: Benchmarking LLMs on Professional Consultation for Real-Life Problems
- Toward Measuring AI's Effects on Skill Formation: The Stock-Formation Gap
Original note title
emergent abilities of LLMs are metric artifacts not fundamental scaling behavior changes