Model Type | text-to-text, decoder-only, large language model |
|
Use Cases |
Primary Use Cases: | Content Creation and Communication - Text Generation, Chatbots and Conversational AI, Research and Education |
|
Limitations: | Training Data Influences - Biases or gaps in training data affect responses, Context and Task Complexity - Challenging for open-ended tasks, Language Ambiguity - Struggles with nuances and figurative language, Factual Accuracy - May generate incorrect or outdated information, Common Sense - Lacks common sense reasoning |
|
Considerations: | Developers advised to use responsibly |
|
|
Additional Notes | Benefits include high-performance open model implementations for Responsible AI development. |
|
Supported Languages | |
Training Details |
Data Sources: | Web Documents, Code, Mathematics |
|
Data Volume: | |
Hardware Used: | |
|
Safety Evaluation |
Methodologies: | Red-teaming, Ethics and safety benchmarks |
|
Risk Categories: | Text-to-Text Content Safety, Text-to-Text Representational Harms, Memorization, Large-scale harm |
|
|
Responsible Ai Considerations |
Fairness: | Models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported. |
|
Transparency: | Model card summarizes details on models' capabilities, limitations, and evaluation. |
|
Accountability: | Encouraged to perform continuous monitoring and report misuse. |
|
Mitigation Strategies: | Content safety guidelines; Monitoring and de-biasing techniques recommended. |
|
|
Input Output |
Input Format: | Text string (e.g., question, prompt, document). |
|
Output Format: | Generated English-language text. |
|
|