Search Results - "Jumelet, Jaap"

Structural Persistence in Language Models: Priming as a Window into Abstract Language Representations by Sinclair, Arabella, Jumelet, Jaap, Zuidema, Willem, Fernández, Raquel

“…We investigate the extent to which modern neural language models are susceptible to structural priming, the phenomenon whereby the structure of a sentence…”
Get full text

diagNNose: A Library for Neural Activation Analysis by Jumelet, Jaap

“…In this paper we introduce diagNNose, an open source library for analysing the activations of deep neural networks. diagNNose contains a wide array of…”
Get full text

Transparency at the Source: Evaluating and Interpreting Language Models With Access to the True Distribution by Jumelet, Jaap, Zuidema, Willem

“…We present a setup for training, evaluating and interpreting neural language models, that uses artificial, language-like data. The data is generated using a…”
Get full text

Feature Interactions Reveal Linguistic Structure in Language Models by Jumelet, Jaap, Zuidema, Willem

“…We study feature interactions in the context of feature attribution methods for post-hoc interpretability. In interpretability research, getting to grips with…”
Get full text

Do Language Models Exhibit Human-like Structural Priming Effects? by Jumelet, Jaap, Zuidema, Willem, Sinclair, Arabella

“…We explore which linguistic factors -- at the sentence and token level -- play an important role in influencing language model predictions, and investigate…”
Get full text

Black Big Boxes: Do Language Models Hide a Theory of Adjective Order? by Jumelet, Jaap, Bylinina, Lisa, Zuidema, Willem, Szymanik, Jakub

“…In English and other languages, multiple adjectives in a complex noun phrase show intricate ordering patterns that have been a target of much linguistic…”
Get full text

Interpretability of Language Models via Task Spaces by Weber, Lucas, Jumelet, Jaap, Bruni, Elia, Hupkes, Dieuwke

“…The usual way to interpret language models (LMs) is to test their performance on different benchmarks and subsequently infer their internal processes. In this…”
Get full text

Attribution and Alignment: Effects of Local Context Repetition on Utterance Production and Comprehension in Dialogue by Molnar, Aron, Jumelet, Jaap, Giulianelli, Mario, Sinclair, Arabella

“…Language models are often used as the backbone of modern dialogue systems. These models are pre-trained on large amounts of written fluent language. Repetition…”
Get full text

DecoderLens: Layerwise Interpretation of Encoder-Decoder Transformers by Langedijk, Anna, Mohebbi, Hosein, Sarti, Gabriele, Zuidema, Willem, Jumelet, Jaap

“…In recent years, many interpretability methods have been proposed to help interpret the internal states of Transformer-models, at different levels of precision…”
Get full text

Curriculum Learning with Adam: The Devil Is in the Wrong Details by Weber, Lucas, Jumelet, Jaap, Michel, Paul, Bruni, Elia, Hupkes, Dieuwke

“…Curriculum learning (CL) posits that machine learning models -- similar to humans -- may learn more efficiently from data that match their current learning…”
Get full text

Do Language Models Understand Anything? On the Ability of LSTMs to Understand Negative Polarity Items by Jumelet, Jaap, Hupkes, Dieuwke

“…In this paper, we attempt to link the inner workings of a neural language model to linguistic theory, focusing on a complex phenomenon well discussed in formal…”
Get full text

The Birth of Bias: A case study on the evolution of gender bias in an English language model by van der Wal, Oskar, Jumelet, Jaap, Schulz, Katrin, Zuidema, Willem

“…Detecting and mitigating harmful biases in modern language models are widely recognized as crucial, open problems. In this paper, we take a step back and…”
Get full text

ChapGTP, ILLC's Attempt at Raising a BabyLM: Improving Data Efficiency by Automatic Task Formation by Jumelet, Jaap, Hanna, Michael, Kloots, Marianne de Heer, Langedijk, Anna, Pouw, Charlotte, van der Wal, Oskar

“…We present the submission of the ILLC at the University of Amsterdam to the BabyLM challenge (Warstadt et al., 2023), in the strict-small track. Our final…”
Get full text

Filtered Corpus Training (FiCT) Shows that Language Models can Generalize from Indirect Evidence by Patil, Abhinav, Jumelet, Jaap, Chiu, Yu Ying, Lapastora, Andy, Shen, Peter, Wang, Lexie, Willrich, Clevis, Steinert-Threlkeld, Shane

“…This paper introduces Filtered Corpus Training, a method that trains language models (LMs) on corpora with certain linguistic constructions filtered out from…”
Get full text

Structural Persistence in Language Models: Priming as a Window into Abstract Language Representations by Sinclair, Arabella, Jumelet, Jaap, Zuidema, Willem, Fernández, Raquel

“…We investigate the extent to which modern, neural language models are susceptible to structural priming, the phenomenon whereby the structure of a sentence…”
Get full text

Analysing Neural Language Models: Contextual Decomposition Reveals Default Reasoning in Number and Gender Assignment by Jumelet, Jaap, Zuidema, Willem, Hupkes, Dieuwke

“…Extensive research has recently shown that recurrent neural language models are able to process a wide range of grammatical phenomena. How these models are…”
Get full text

Attention vs non-attention for a Shapley-based explanation method by Kersten, Tom, Wong, Hugh Mee, Jumelet, Jaap, Hupkes, Dieuwke

“…The field of explainable AI has recently seen an explosion in the number of explanation methods for highly non-linear deep neural networks. The extent to which…”
Get full text

Language Modelling as a Multi-Task Problem by Weber, Lucas, Jumelet, Jaap, Bruni, Elia, Hupkes, Dieuwke

“…In this paper, we propose to study language modelling as a multi-task problem, bringing together three strands of research: multi-task learning, linguistics,…”
Get full text

Language Models Use Monotonicity to Assess NPI Licensing by Jumelet, Jaap, Denić, Milica, Szymanik, Jakub, Hupkes, Dieuwke, Steinert-Threlkeld, Shane

“…We investigate the semantic knowledge of language models (LMs), focusing on (1) whether these LMs create categories of linguistic environments based on their…”
Get full text