Factual associations in autoregressive transformers are localized to mid-layer feed-forward modules and can be edited via rank-one model editing while preserving both specificity and generalization on counterfactual tests.
Measuring and Improving Consistency in Pretrained Language Models
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
representative citing papers
Proposes equation-grounded taxonomy (unexpected AIS activity, route deviation, close approach) and LLM-guided synthesis pipeline to generate timestamp-labeled anomalies for evaluating maritime detection models.
LLMs display inconsistent factual recall across different surface forms of the same entity, with greater robustness to minor spelling changes than to aliases or abbreviations.
LMs store facts in task-specific parameter subsets, shown by inconsistent emergence across tasks during training and distinct localized parameters for the same fact.
LLM 2-bit quantization fails via either cumulative signal degradation or early computation collapse in key components.
citing papers explorer
-
Locating and Editing Factual Associations in GPT
Factual associations in autoregressive transformers are localized to mid-layer feed-forward modules and can be edited via rank-one model editing while preserving both specificity and generalization on counterfactual tests.
-
Redefining Maritime Anomaly Detection via Equation-Grounded Synthetic Anomalies
Proposes equation-grounded taxonomy (unexpected AIS activity, route deviation, close approach) and LLM-guided synthesis pipeline to generate timestamp-labeled anomalies for evaluating maritime detection models.
-
Revisiting Non-Verbatim Memorization in Large Language Models: The Role of Entity Surface Forms
LLMs display inconsistent factual recall across different surface forms of the same entity, with greater robustness to minor spelling changes than to aliases or abbreviations.
-
LMs as Task-Specific Knowledge Bases: An Interpretability Analysis
LMs store facts in task-specific parameter subsets, shown by inconsistent emergence across tasks during training and distinct localized parameters for the same fact.
-
From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization
LLM 2-bit quantization fails via either cumulative signal degradation or early computation collapse in key components.