paper-with-me

홈 › Papers

Underspecification in Language Modeling Tasks: A Causality-Informed Study of Gendered Pronoun Resolution

2022-09-30 · Emily McMilin

Modern language modeling tasks are often underspecified: for a given token prediction, many words may satisfy the user's intent of producing natural language at inference time, however only one word will minimize the task's loss function at training time. We introduce a simple causal mechanism to describe the role underspecification plays in the generation of spurious correlations. Despite its simplicity, our causal model directly informs the development of two lightweight black-box evaluation methods, that we apply to gendered pronoun resolution tasks on a wide range of LLMs to 1) aid in the detection of inference-time task underspecification by exploiting 2) previously unreported gender vs. time and gender vs. location spurious correlations on LLMs with a range of A) sizes: from BERT-base to GPT-4 Turbo Preview, B) pre-training objectives: from masked & autoregressive language modeling to a mixture of these objectives, and C) training stages: from pre-training only to reinforcement learning from human feedback (RLHF). Code and open-source demos available at https://github.com/2dot71mily/uspec.

📄 PDF Abstract BibTeX arXiv:2210.00131

Code (2)

2dot71mily/sib_paper 공식 구현
2dot71mily/uspec 공식 구현

Tasks

Language ModelingLanguage ModellingSelection bias

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Attention Dropout Attention Dropout is a type of dropout used in attention-based architectures, where elements are randomly dropped out of the…
Weight Decay 설명 없음
Discriminative Fine-Tuning Discriminative Fine-Tuning is a fine-tuning strategy that is used for ULMFiT type models. Instead of using the same learning rate…
Residual Connection 설명 없음

Similar Papers 제목 키워드 기반

Underspecification Presents Challenges for Credibility in Modern Machine Learning

2020-11-06 · Alexander D'Amour, Katherine Heller, Dan Moldovan, Ben Adlam 외

ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains. We identify underspecification as a key reason for these failures. An ML pipeline is underspecified when it can return many…

BIG-bench Machine Learning

Accounting for Underspecification in Statistical Claims of Model Superiority

2025-11-04 · Thomas Sanchez, Pedro M. Gordaliza, Meritxell Bach Cuadra arxiv

Machine learning methods are increasingly applied in medical imaging, yet many reported improvements lack statistical robustness: recent works have highlighted that small but significant performance gains are highly like…

A Residual Guided strategy with Generative Adversarial Networks in training Physics-Informed Transformer Networks

2025-07-15 · Ziyang Zhang, Feifan Zhang, Weidong Tang, Lei Shi 외 arxiv

Nonlinear partial differential equations (PDEs) are pivotal in modeling complex physical systems, yet traditional Physics-Informed Neural Networks (PINNs) often struggle with unresolved residuals in critical spatiotempor…

Causality for Large Language Models

2024-10-20 · Anpeng Wu, Kun Kuang, Minqin Zhu, Yingrong Wang 외

Recent breakthroughs in artificial intelligence have driven a paradigm shift, where large language models (LLMs) with billions or trillions of parameters are trained on vast datasets, achieving unprecedented success acro…

Prompt Engineering

Dealing with Semantic Underspecification in Multimodal NLP

2023-06-08 · Sandro Pezzelle

Intelligent systems that aim at mastering language as humans do must deal with its semantic underspecification, namely, the possibility for a linguistic signal to convey only part of the information needed for communicat…