Pith. sign in

REVIEW 13 cited by

Overtrained Language Models Are Harder to Fine-Tune

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 2503.19206 v2 pith:VDYYBT3S submitted 2025-03-24 cs.CL cs.AI

Overtrained Language Models Are Harder to Fine-Tune

classification cs.CL cs.AI
keywords modelsperformancepre-trainedpre-trainingassumptioncatastrophicdownstreamfine-tune
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Large language models are pre-trained on ever-growing token budgets under the assumption that better pre-training performance translates to improved downstream models. In this work, we challenge this assumption and show that extended pre-training can make models harder to fine-tune, leading to degraded final performance. We term this phenomenon catastrophic overtraining. For example, the instruction-tuned OLMo-1B model pre-trained on 3T tokens leads to over 2% worse performance on multiple standard LLM benchmarks than its 2.3T token counterpart. Through controlled experiments and theoretical analysis, we show that catastrophic overtraining arises from a systematic increase in the broad sensitivity of pre-trained parameters to modifications, including but not limited to fine-tuning. Our findings call for a critical reassessment of pre-training design that considers the downstream adaptability of the model.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 13 Pith papers

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

  1. Bridging Compute- and Data-Optimal Pretraining

    cs.LG 2026-07 conditional novelty 7.0

    Pretraining loss obeys a single law in which repeated or paraphrased tokens count as η(N, data-per-parameter, expansion-ratio) fresh tokens, with total effective data saturating as derived tokens grow.

  2. Understanding Reasoning from Pretraining to Post-Training

    cs.LG 2026-07 conditional novelty 7.0

    A joint scaling law: post-RL chess and math performance is predictable from pretraining loss, RL improvement rate grows with pretraining tokens, and RL both amplifies and discovers moves depending on difficulty.

  3. Disentangling Geometry, Performance, and Training in Language Models

    cs.CL 2026-02 conditional novelty 7.0

    Effective rank of the unembedding matrix mainly reflects hyperparameters like batch size and weight decay and is not a reliable predictor of language-model performance.

  4. Improving Neural Network Training by Decoupling the Magnitude and Direction of Weight Vectors

    cs.LG 2026-06 unverdicted novelty 6.0

    MD Decoupling factorizes weights into fixed-norm directions and learnable per-row/column magnitudes updated at independent rates, improving Adam and Muon training stability and scale transfer without weight decay or warmup.

  5. LLMs as Noisy Channels: A Shannon Perspective on Model Capacity and Scaling Laws

    cs.LG 2026-05 unverdicted novelty 6.0

    The Shannon Scaling Law treats LLM training as noisy-channel transmission and predicts U-shaped performance degradation when signal-to-noise ratio falls below a threshold, outperforming monotonic scaling laws on Pythi...

  6. Early Data Exposure Improves Robustness to Subsequent Fine-Tuning

    cs.LG 2026-05 conditional novelty 6.0

    Early mixing of post-training data into pretraining improves retention of acquired capabilities after subsequent fine-tuning in language models.

  7. Optimizer-Model Consistency: Full Finetuning with the Same Optimizer as Pretraining Forgets Less

    cs.LG 2026-05 unverdicted novelty 6.0

    Full finetuning with the pretraining optimizer reduces forgetting compared to other optimizers or LoRA while achieving comparable new-task performance.

  8. Nexus: Same Pretraining Loss, Better Downstream Generalization via Common Minima

    cs.LG 2026-04 unverdicted novelty 6.0

    Nexus optimizer improves LLM downstream performance by converging to common minima across data sources despite identical pretraining loss.

  9. Nexus: Same Pretraining Loss, Better Downstream Generalization via Common Minima

    cs.LG 2026-04 unverdicted novelty 6.0

    Nexus matches standard pretraining loss while improving downstream and OOD performance by maximizing gradient similarity so task-specific minima stay close.

  10. Manifold-Constrained Hyper-Connections for Parameter-Efficient Finetuning

    cs.LG 2026-07 conditional novelty 5.0

    Applying mHC as a PEFT method shows that learned residual mixing is unnecessary — even harmful — in finetuning, and mHC+LoRA combinations give small task-dependent gains.

  11. Improving Neural Network Training by Decoupling the Magnitude and Direction of Weight Vectors

    cs.LG 2026-06 conditional novelty 5.0

    Splitting weight matrices into a fixed-norm direction and learnable per-row/column magnitudes improves LLM training over AdamW/Muon, removes weight decay and warmup, and transfers the optimal LR across width.

  12. Forgetting in Language Models: Capacity, Optimization, and Self-Generated Replay

    cs.LG 2026-05 unverdicted novelty 5.0

    Self-generated replay from language models nearly eliminates catastrophic forgetting during finetuning except when models are pretrained close to saturation.

  13. Reward-Free Code Alignment from Pretrained or Fine-Tuned LLM: Unpacking the Trade-offs for Code Generation

    cs.SE 2026-06 unverdicted novelty 4.0

    Empirical study on five LLMs finds pretrained-to-aligned paths yield bigger gains over baseline than finetuned-to-aligned paths, though absolute accuracy remains lower for pretrained starts.