|
<< Click to Display Table of Contents >> Navigation: 3. Script Language > AI - Artificial Intelligence Commands > AIG. - Google AI > 4. Configuration (Generation Parameters) > AIG. - AI Google Gemini Integration |
MiniRobotLanguage (MRL)
AIG.Get TopP
Retrieve Current Top P (Nucleus Sampling) Value
Intention
GetTopP Command: Checking Response Diversity Setting
The GetTopP Command retrieves the current Top P (nucleus sampling) value used by the Google Generative AI API to control response diversity.
This helps you verify or adjust your script based on the AI�s output behavior.
It�s part of the Generation Parameters suite within AIG.
The GetTopP Command fetches the current value of the AIG_TopP global variable, which sets the nucleus sampling threshold for AI responses.
You can store this value in a variable or, if no variable is specified, it will be placed on the Top of Stack (TOS).
Retrieving the Top P value is useful for:
�Verification: Confirm the setting applied by AIG.Set TopP.
�Script Logic: Adjust behavior based on the diversity level.
�Debugging: Ensure the AI�s output aligns with your expectations.
Call this Command to retrieve the current Top P value, optionally storing it in a variable.
The default is 0.95 unless modified by AIG.Set TopP.
Example Usage
AIG.Get TopP|$$TOP
DBP.Current Top P: $$TOP
AIG.Get TopP
POP.$$TOP
DBP.Top P from Stack: $$TOP
The returned value reflects the last set Top P or the default 0.95.
Illustration
+---------------+
� Top P � � � � �
+---------------�
� 0.7 � � � � � �
+---------------+
Example output after setting Top P to 0.7.
Syntax
AIG.GetTopP[|P1]
AIG.Get TopP[|P1]
Parameter Explanation
P1 - (Optional) The variable where the Top P value will be stored. If omitted, the value is placed on the Top of Stack (TOS).
Example
AIG.Set TopP|0.4
AIG.Get TopP|$$TOP
DBP.Current Top P: $$TOP
ENR.
Remarks
- The default Top P is 0.95 if not previously set.
- Use this Command after AIG.Set TopP to confirm the setting.
Limitations
- Only reflects the current session�s setting or default.
- Does not indicate if the value is optimal for the chosen model.
See also: