Picture for Roger Levy

Roger Levy

MIT

Elements of World Knowledge (EWOK): A cognition-inspired framework for evaluating basic world knowledge in language models

Add code
May 15, 2024
Viaarxiv icon

Learning Phonotactics from Linguistic Informants

May 08, 2024
Figure 1 for Learning Phonotactics from Linguistic Informants
Figure 2 for Learning Phonotactics from Linguistic Informants
Figure 3 for Learning Phonotactics from Linguistic Informants
Figure 4 for Learning Phonotactics from Linguistic Informants
Viaarxiv icon

Language models align with human judgments on key grammatical constructions

Add code
Jan 19, 2024
Viaarxiv icon

LINC: A Neurosymbolic Approach for Logical Reasoning by Combining Language Models with First-Order Logic Provers

Add code
Oct 23, 2023
Viaarxiv icon

Probing self-supervised speech models for phonetic and phonemic information: a case study in aspiration

Jun 09, 2023
Figure 1 for Probing self-supervised speech models for phonetic and phonemic information: a case study in aspiration
Figure 2 for Probing self-supervised speech models for phonetic and phonemic information: a case study in aspiration
Figure 3 for Probing self-supervised speech models for phonetic and phonemic information: a case study in aspiration
Figure 4 for Probing self-supervised speech models for phonetic and phonemic information: a case study in aspiration
Viaarxiv icon

A Cross-Linguistic Pressure for Uniform Information Density in Word Order

Add code
Jun 06, 2023
Figure 1 for A Cross-Linguistic Pressure for Uniform Information Density in Word Order
Figure 2 for A Cross-Linguistic Pressure for Uniform Information Density in Word Order
Figure 3 for A Cross-Linguistic Pressure for Uniform Information Density in Word Order
Figure 4 for A Cross-Linguistic Pressure for Uniform Information Density in Word Order
Viaarxiv icon

Prompt-based methods may underestimate large language models' linguistic generalizations

Add code
May 22, 2023
Figure 1 for Prompt-based methods may underestimate large language models' linguistic generalizations
Figure 2 for Prompt-based methods may underestimate large language models' linguistic generalizations
Figure 3 for Prompt-based methods may underestimate large language models' linguistic generalizations
Figure 4 for Prompt-based methods may underestimate large language models' linguistic generalizations
Viaarxiv icon

The neural dynamics of auditory word recognition and integration

May 22, 2023
Figure 1 for The neural dynamics of auditory word recognition and integration
Figure 2 for The neural dynamics of auditory word recognition and integration
Figure 3 for The neural dynamics of auditory word recognition and integration
Figure 4 for The neural dynamics of auditory word recognition and integration
Viaarxiv icon

Expectations over Unspoken Alternatives Predict Pragmatic Inferences

Add code
Apr 07, 2023
Figure 1 for Expectations over Unspoken Alternatives Predict Pragmatic Inferences
Figure 2 for Expectations over Unspoken Alternatives Predict Pragmatic Inferences
Figure 3 for Expectations over Unspoken Alternatives Predict Pragmatic Inferences
Figure 4 for Expectations over Unspoken Alternatives Predict Pragmatic Inferences
Viaarxiv icon

Language model acceptability judgements are not always robust to context

Dec 18, 2022
Figure 1 for Language model acceptability judgements are not always robust to context
Figure 2 for Language model acceptability judgements are not always robust to context
Figure 3 for Language model acceptability judgements are not always robust to context
Figure 4 for Language model acceptability judgements are not always robust to context
Viaarxiv icon