Pith. sign in

REVIEW 4 cited by

A Tale of Tails: Model Collapse as a Change of Scaling Laws

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 2402.07043 v2 pith:JYMI52FY submitted 2024-02-10 cs.LG cs.AIcs.CL

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

As AI model size grows, neural scaling laws have become a crucial tool to predict the improvements of large models when increasing capacity and the size of original (human or natural) training data. Yet, the widespread use of popular models means that the ecosystem of online data and text will co-evolve to progressively contain increased amounts of synthesized data. In this paper we ask: How will the scaling laws change in the inevitable regime where synthetic data makes its way into the training corpus? Will future models, still improve, or be doomed to degenerate up to total (model) collapse? We develop a theoretical framework of model collapse through the lens of scaling laws. We discover a wide range of decay phenomena, analyzing loss of scaling, shifted scaling with number of generations, the ''un-learning" of skills, and grokking when mixing human and synthesized data. Our theory is validated by large-scale experiments with a transformer on an arithmetic task and text generation using the large language model Llama2.

Discussion (0). Sign in to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 39 citations worldwide. Full citation record

  1. What Matters in LLM-generated Data: Diversity and Its Effect on Model Fine-Tuning

    cs.CL 2025-06 conditional novelty 6.0 of 10

    Moderately diverse LLM-generated data can improve fine-tuned model performance in low-data settings when distribution shift is minimal, while high diversity or large distribution shift hurts.

  2. Epistemic diversity across language models mitigates knowledge collapse

    cs.LG 2025-12 reject novelty 5.0 of 10

    In repeated self-training loops on Wikitext2, ecosystems of four small language models show lower average perplexity than one, two, or sixteen models, but the paper's broader claims about monotonic optima, robustness,...

  3. Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training

    cs.AI 2025-05 conditional novelty 5.0 of 10

    Reinforcing the two experts most correlated with thinking tokens improves reasoning accuracy and efficiency in MoE large reasoning models, with gains of up to 10 points on AIME benchmarks.

  4. Securing High-Concurrency Ticket Sales: A Framework Based on Microservice

    cs.SE 2025-12 reject novelty 2.0 of 10

    A railway ticketing system built from standard Spring Cloud components is reported to reach 817 req/s on a train-query interface, but only under a 100-thread local VM test with inconsistent purchase-interface data.

Pith tools