Picture for Willem Zuidema

Willem Zuidema

Enforcing Interpretability in Time Series Transformers: A Concept Bottleneck Framework

Add code
Oct 08, 2024
Viaarxiv icon

Disentangling Textual and Acoustic Features of Neural Speech Representations

Add code
Oct 03, 2024
Figure 1 for Disentangling Textual and Acoustic Features of Neural Speech Representations
Figure 2 for Disentangling Textual and Acoustic Features of Neural Speech Representations
Figure 3 for Disentangling Textual and Acoustic Features of Neural Speech Representations
Figure 4 for Disentangling Textual and Acoustic Features of Neural Speech Representations
Viaarxiv icon

Human-like Linguistic Biases in Neural Speech Models: Phonetic Categorization and Phonotactic Constraints in Wav2Vec2.0

Add code
Jul 03, 2024
Viaarxiv icon

Black Big Boxes: Do Language Models Hide a Theory of Adjective Order?

Add code
Jul 02, 2024
Viaarxiv icon

Perception of Phonological Assimilation by Neural Speech Recognition Models

Add code
Jun 21, 2024
Viaarxiv icon

Do Language Models Exhibit Human-like Structural Priming Effects?

Add code
Jun 07, 2024
Viaarxiv icon

Transparency at the Source: Evaluating and Interpreting Language Models With Access to the True Distribution

Add code
Oct 23, 2023
Viaarxiv icon

Identifying and Adapting Transformer-Components Responsible for Gender Bias in an English Language Model

Add code
Oct 19, 2023
Viaarxiv icon

Homophone Disambiguation Reveals Patterns of Context Mixing in Speech Transformers

Add code
Oct 15, 2023
Viaarxiv icon

DecoderLens: Layerwise Interpretation of Encoder-Decoder Transformers

Add code
Oct 05, 2023
Viaarxiv icon