Pith. sign in

REVIEW 3 cited by

Quantifying Generalization Complexity for Large Language Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.01769 v2 pith:CCWHYHTL submitted 2024-10-02 cs.CL

classification cs.CL
keywords complexitygeneralizationllmsmodelscriticalcapabilitiesevaluationmemorization
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

While large language models (LLMs) have shown exceptional capabilities in understanding complex queries and performing sophisticated tasks, their generalization abilities are often deeply entangled with memorization, necessitating more precise evaluation. To address this challenge, we introduce Scylla, a dynamic evaluation framework that quantitatively measures the generalization abilities of LLMs. Scylla disentangles generalization from memorization via assessing model performance on both in-distribution (ID) and out-of-distribution (OOD) data through 20 tasks across 5 levels of complexity. Through extensive experiments, we uncover a non-monotonic relationship between task complexity and the performance gap between ID and OOD data, which we term the generalization valley. Specifically, this phenomenon reveals a critical threshold - referred to as critical complexity - where reliance on non-generalizable behavior peaks, indicating the upper bound of LLMs' generalization capabilities. As model size increases, the critical complexity shifts toward higher levels of task complexity, suggesting that larger models can handle more complex reasoning tasks before over-relying on memorization. Leveraging Scylla and the concept of critical complexity, we benchmark 28LLMs including both open-sourced models such as LLaMA and Qwen families, and close-sourced models like Claude and GPT, providing a more robust evaluation and establishing a clearer understanding of LLMs' generalization capabilities.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Enough is as good as a feast: A Comprehensive Analysis of How Reinforcement Learning Mitigates Task Conflicts in LLMs

    cs.CL 2026-07 conditional novelty 6.0 of 10

    RL-trained LLMs keep most of their skills after weight merging, while SFT-trained LLMs drop about 19% on average, because RL keeps parameter updates smaller and more task-compatible.

  2. Mitigating Object Hallucination via Robust Local Perception Search

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A training-free decoding method that uses an MLLM's own local object descriptions as a reward prior, combined with CLIP similarity, to cut object hallucination, especially under adversarial image noise.

  3. LLM Evaluation Based on Aerospace Manufacturing Expertise: Automated Generation and Multi-Model Question Answering

    cs.CL 2025-01 conditional novelty 6.0 of 10

    The authors created an aerospace manufacturing QA benchmark with roughly 2,480 questions and found that top LLMs score below 51% accuracy.

Pith tools