Pith. sign in

REVIEW 5 cited by

Preserving Pre-trained Features Helps Calibrate Fine-tuned 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 2305.19249 v1 pith:YCIJQFQT submitted 2023-05-30 cs.CL cs.LG

classification cs.CLcs.LG
keywords languagefine-tunedmodelscalibrationpre-trainedfeaturesdownstreammethods
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Large pre-trained language models (PLMs) have demonstrated strong performance on natural language understanding (NLU) tasks through fine-tuning. However, fine-tuned models still suffer from overconfident predictions, especially in out-of-domain settings. In this paper, we tackle the problem of calibrating fine-tuned language models. We demonstrate that the PLMs are well-calibrated on the masked language modeling task with robust predictive confidence under domain shift, yet the fine-tuned models fail to retain such property due to catastrophic forgetting, which impacts the calibration on the downstream classification task. In light of these observations, we evaluate the calibration of several methods that preserve pre-trained features and show that preserving pre-trained features can improve the calibration of fine-tuned language models. Among these methods, our proposed method that encourages the fine-tuned model to learn generative representations with auxiliary language modeling objective achieves competitive accuracy and the lowest expected calibration error compared to several strong baselines under both in-domain and out-of-domain settings on three downstream NLU tasks.

Discussion (0). Sign in to comment.

Forward citations

Cited by 5 Pith papers

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

  1. Rationales Are Not Silver Bullets: Measuring the Impact of Rationales on Model Performance and Reliability

    cs.CL 2025-05 conditional novelty 7.0 of 10

    Rationale-augmented finetuning can hurt accuracy while improving calibration, with the sizes of both effects tied linearly to task difficulty.

  2. TabPFN beyond Tabular Data: Calibration and Accuracy on Multimodal Embeddings

    cs.LG 2026-07 conditional novelty 6.0 of 10

    TabPFN as a zero-gradient head on frozen multimodal embeddings ranks best on NLL and ECE across 22 820 episodes while matching accuracy in mid-shot, mid-dimension regimes and also fixes miscalibration after fine-tuning.

  3. From Calibration to Collaboration: LLM Uncertainty Quantification Should Be More Human-Centered

    cs.CL 2025-06 conditional novelty 5.0 of 10

    LLM uncertainty quantification should be judged by whether it improves real human decisions, not by calibration scores on trivia benchmarks.

  4. Optimising Language Models for Downstream Tasks: A Post-Training Perspective

    cs.CL 2025-06 conditional novelty 3.0 of 10

    A dissertation that repackages the author's previously published papers on continued pre-training, prompt tuning, and instruction modelling into a single narrative.

  5. Loki's Dance of Illusions: A Comprehensive Survey of Hallucination in Large Language Models

    cs.CL 2025-06 reject novelty 3.0 of 10

    A survey of LLM hallucination research that formalizes hallucination types and argues, via incompleteness and undecidability arguments, that hallucinations cannot be fully eliminated.

Pith tools