Linear Recency Bias During Training Improves Transformers' Fit to Reading Times
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929502288871424 |
|---|---|
| author | Clark, Christian Oh, Byung-Doh Schuler, William |
| author_facet | Clark, Christian Oh, Byung-Doh Schuler, William |
| contents | Recent psycholinguistic research has compared human reading times to surprisal estimates from language models to study the factors shaping human sentence processing difficulty. Previous studies have shown a strong fit between surprisal values from Transformers and reading times. However, standard Transformers work with a lossless representation of the entire previous linguistic context, unlike models of human language processing that include memory decay. To bridge this gap, this paper evaluates a modification of the Transformer model that uses ALiBi (Press et al., 2022), a recency bias added to attention scores. Surprisal estimates with ALiBi show an improved fit to human reading times compared to a standard Transformer baseline. A subsequent analysis of attention heads suggests that ALiBi's mixture of slopes -- which determine the rate of memory decay in each attention head -- may play a role in the improvement by helping models with ALiBi to track different kinds of linguistic dependencies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_11250 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Linear Recency Bias During Training Improves Transformers' Fit to Reading Times Clark, Christian Oh, Byung-Doh Schuler, William Computation and Language Recent psycholinguistic research has compared human reading times to surprisal estimates from language models to study the factors shaping human sentence processing difficulty. Previous studies have shown a strong fit between surprisal values from Transformers and reading times. However, standard Transformers work with a lossless representation of the entire previous linguistic context, unlike models of human language processing that include memory decay. To bridge this gap, this paper evaluates a modification of the Transformer model that uses ALiBi (Press et al., 2022), a recency bias added to attention scores. Surprisal estimates with ALiBi show an improved fit to human reading times compared to a standard Transformer baseline. A subsequent analysis of attention heads suggests that ALiBi's mixture of slopes -- which determine the rate of memory decay in each attention head -- may play a role in the improvement by helping models with ALiBi to track different kinds of linguistic dependencies. |
| title | Linear Recency Bias During Training Improves Transformers' Fit to Reading Times |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2409.11250 |