agent_inspect.models package

Subpackages

• agent_inspect.models.metrics package
• agent_inspect.models.tools package
• agent_inspect.models.user_proxy package

Submodules

agent_inspect.models.llm_payload module

class agent_inspect.models.llm_payload.LLMPayload(user_prompt, model=None, system_prompt=None, temperature=None, max_tokens=None, structured_output=None)

Bases: object

Represents the payload to be sent to a Language Model (LLM) for processing.

Parameters:

• user_prompt (str)

• model (str | None)

• system_prompt (str | None)

• temperature (float | None)

• max_tokens (int | None)

• structured_output (Any | None)

max_tokens: Optional[int] = None

The maximum number of tokens to be generated in the LLM’s response.

model: Optional[str] = None

The specific LLM model to be used for processing the prompt.

structured_output: Optional[Any] = None

An optional structured format for the LLM’s output, if applicable.

system_prompt: Optional[str] = None

The system-level prompt that provides context or instructions to the LLM.

temperature: Optional[float] = None

The temperature setting for the LLM, influencing the randomness of its output.

user_prompt: str

The raw text prompt provided by the user to the LLM for processing.

agent_inspect.models.llm_response module

class agent_inspect.models.llm_response.LLMResponse(status, completion=None, error_message=None)

Bases: object

Represents a response from a LLM.

Parameters:

• status (int)

• completion (str | None)

• error_message (str | None)

completion: Optional[str] = None

Represents the text output generated by the LLM in response to an input prompt. Can be none if there is no output generated from the LLM due to an error.

error_message: Optional[str] = None

Contains error related information from the LLM in response to an input prompt. Can be none if there is no error occurred.

status: int

Contains the http status code from the LLM.

Module contents

class agent_inspect.models.LLMPayload(user_prompt, model=None, system_prompt=None, temperature=None, max_tokens=None, structured_output=None)

Bases: object

Represents the payload to be sent to a Language Model (LLM) for processing.

Parameters:

• user_prompt (str)

• model (str | None)

• system_prompt (str | None)

• temperature (float | None)

• max_tokens (int | None)

• structured_output (Any | None)

max_tokens: Optional[int] = None

The maximum number of tokens to be generated in the LLM’s response.

model: Optional[str] = None

The specific LLM model to be used for processing the prompt.

structured_output: Optional[Any] = None

An optional structured format for the LLM’s output, if applicable.

system_prompt: Optional[str] = None

The system-level prompt that provides context or instructions to the LLM.

temperature: Optional[float] = None

The temperature setting for the LLM, influencing the randomness of its output.

user_prompt: str

The raw text prompt provided by the user to the LLM for processing.

class agent_inspect.models.LLMResponse(status, completion=None, error_message=None)

Bases: object

Represents a response from a LLM.

Parameters:

• status (int)

• completion (str | None)

• error_message (str | None)

completion: Optional[str] = None

Represents the text output generated by the LLM in response to an input prompt. Can be none if there is no output generated from the LLM due to an error.

error_message: Optional[str] = None

Contains error related information from the LLM in response to an input prompt. Can be none if there is no error occurred.

status: int

Contains the http status code from the LLM.