Document Analysis
#3
by
parthmm
- opened
README.md
CHANGED
@@ -10,7 +10,6 @@ widget:
|
|
10 |
|
11 |
Note: This model was previously known as PubMedGPT 2.7B, but we have changed it due to a request from the NIH which holds the trademark for "PubMed".
|
12 |
|
13 |
-
Paper: [BioMedLM: A 2.7B Parameter Language Model Trained On Biomedical Text](https://arxiv.org/abs/2403.18421)
|
14 |
|
15 |
BioMedLM 2.7B is new language model trained exclusively on biomedical abstracts and papers from [The Pile](https://pile.eleuther.ai/). This GPT-style model can achieve strong results on a variety of biomedical NLP tasks, including a new state of the art performance of 50.3% accuracy on the MedQA biomedical question answering task.
|
16 |
|
@@ -80,7 +79,7 @@ We do not recommend using this model for natural language generation in a produc
|
|
80 |
# Bias, Risks, and Limitations
|
81 |
|
82 |
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
83 |
-
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
|
84 |
|
85 |
## Recommendations
|
86 |
|
10 |
|
11 |
Note: This model was previously known as PubMedGPT 2.7B, but we have changed it due to a request from the NIH which holds the trademark for "PubMed".
|
12 |
|
|
|
13 |
|
14 |
BioMedLM 2.7B is new language model trained exclusively on biomedical abstracts and papers from [The Pile](https://pile.eleuther.ai/). This GPT-style model can achieve strong results on a variety of biomedical NLP tasks, including a new state of the art performance of 50.3% accuracy on the MedQA biomedical question answering task.
|
15 |
|
79 |
# Bias, Risks, and Limitations
|
80 |
|
81 |
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
82 |
+
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Weidinger et al. (2021)](https://arxiv.org/pdf/2112.04359.pdf)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
|
83 |
|
84 |
## Recommendations
|
85 |
|