| Model Type | | text-to-text, decoder-only, large language model |
|
| Use Cases |
| Areas: | | Content Creation and Communication, Research and Education |
|
| Applications: | | Text Generation, Chatbots and Conversational AI, Text Summarization, NLP Research, Language Learning Tools, Knowledge Exploration |
|
| Limitations: | | Training Data, Context and Task Complexity, Language Ambiguity and Nuance, Factual Accuracy, Common Sense |
|
| Considerations: | | Consider performing continuous monitoring and exploration of de-biasing techniques. |
|
|
| Supported Languages | |
| Training Details |
| Data Sources: | | Web Documents, Code, Mathematics |
|
| Data Volume: | | 9B model trained with 8 trillion tokens |
|
| Methodology: | | Text-to-text, decoder-only large language models |
|
| Hardware Used: | |
|
| Safety Evaluation |
| Methodologies: | | structured evaluations, internal red-teaming testing |
|
| Risk Categories: | | Text-to-Text Content Safety, Text-to-Text Representational Harms, Memorization, Large-scale harm |
|
| Ethical Considerations: | | Child safety, content safety, representational harms, memorization, large-scale harms |
|
|
| Responsible Ai Considerations |
| Fairness: | | Bias and Fairness evaluations against datasets such as WinoBias and BBQ Dataset |
|
| Transparency: | | Model card provides details on architecture, capabilities, limitations, and evaluation processes. |
|
| Accountability: | |
| Mitigation Strategies: | | Continuous monitoring, content safety mechanisms, developer education |
|
|
| Input Output |
| Input Format: | |
| Accepted Modalities: | |
| Output Format: | | Generated English-language text |
|
|
