neuronpedia_inference_client 1.2.0

Neuronpedia - Inference Server

neuronpedia-inference-client

No description provided (generated by Openapi Generator https://github.com/openapitools/openapi-generator)

This Python package is automatically generated by the OpenAPI Generator project:

• API version: 1.1.0
• Package version: 1.2.0
• Generator version: 7.12.0
• Build package: org.openapitools.codegen.languages.PythonClientCodegen

Requirements.

Python 3.8+

Installation & Usage

pip install

If the python package is hosted on a repository, you can install directly using:

pip install git+https://github.com/GIT_USER_ID/GIT_REPO_ID.git

(you may need to run pip with root permission: sudo pip install git+https://github.com/GIT_USER_ID/GIT_REPO_ID.git)

Then import the package:

import neuronpedia_inference_client

Setuptools

Install via Setuptools.

python setup.py install --user

(or sudo python setup.py install to install the package for all users)

Then import the package:

import neuronpedia_inference_client

Tests

Execute pytest to run the tests.

Getting Started

Please follow the installation procedure and then run the following:

import neuronpedia_inference_client
from neuronpedia_inference_client.rest import ApiException
from pprint import pprint

# Defining the host is optional and defaults to /v1
# See configuration.py for a list of all supported configuration parameters.
configuration = neuronpedia_inference_client.Configuration(
    host = "/v1"
)

# The client must configure the authentication and authorization parameters
# in accordance with the API server security policy.
# Examples for each auth method are provided below, use the example that
# satisfies your auth use case.

# Configure API key authorization: SimpleSecretAuth
configuration.api_key['SimpleSecretAuth'] = os.environ["API_KEY"]

# Uncomment below to setup prefix (e.g. Bearer) for API key, if needed
# configuration.api_key_prefix['SimpleSecretAuth'] = 'Bearer'


# Enter a context with an instance of the API client
with neuronpedia_inference_client.ApiClient(configuration) as api_client:
    # Create an instance of the API class
    api_instance = neuronpedia_inference_client.DefaultApi(api_client)
    activation_all_post_request = neuronpedia_inference_client.ActivationAllPostRequest() # ActivationAllPostRequest | 

    try:
        # For a given prompt, get the top activating features for a set of SAEs (eg gemmascope-res-65k), or specific SAEs in the set of SAEs (eg 0-gemmascope-res-65k, 5-gemmascope-res-65k). Also has other customization options.
        api_response = api_instance.activation_all_post(activation_all_post_request)
        print("The response of DefaultApi->activation_all_post:\n")
        pprint(api_response)
    except ApiException as e:
        print("Exception when calling DefaultApi->activation_all_post: %s\n" % e)

Documentation for API Endpoints

All URIs are relative to /v1

Class	Method	HTTP request	Description
DefaultApi	activation_all_post	POST /activation/all	For a given prompt, get the top activating features for a set of SAEs (eg gemmascope-res-65k), or specific SAEs in the set of SAEs (eg 0-gemmascope-res-65k, 5-gemmascope-res-65k). Also has other customization options.
DefaultApi	activation_single_post	POST /activation/single	Given a text prompt, returns the activation values for a single SAE latent or custom vector+hook.
DefaultApi	activation_topk_by_token_post	POST /activation/topk-by-token	For a given prompt, get the top activating features at each token position for a single SAE.
DefaultApi	steer_completion_chat_post	POST /steer/completion-chat	For a given prompt, complete it by steering with the given feature or vector
DefaultApi	steer_completion_post	POST /steer/completion	For a given prompt, complete it by steering with the given feature or vector
DefaultApi	tokenize_post	POST /tokenize	Tokenize input text for a given model
DefaultApi	util_sae_topk_by_decoder_cossim_post	POST /util/sae-topk-by-decoder-cossim	Given a specific vector or SAE feature, return the top features by cosine similarity in the same SAE
DefaultApi	util_sae_vector_post	POST /util/sae-vector	Get the raw vector for an SAE feature

Documentation For Models

• ActivationAllPost200Response
• ActivationAllPost200ResponseActivationsInner
• ActivationAllPostRequest
• ActivationSinglePost200Response
• ActivationSinglePost200ResponseActivation
• ActivationSinglePostRequest
• ActivationTopkByTokenPost200Response
• ActivationTopkByTokenPost200ResponseResultsInner
• ActivationTopkByTokenPost200ResponseResultsInnerTopFeaturesInner
• ActivationTopkByTokenPostRequest
• NPFeature
• NPSteerChatMessage
• NPSteerChatResult
• NPSteerCompletionResponseInner
• NPSteerFeature
• NPSteerMethod
• NPSteerType
• NPSteerVector
• SteerCompletionChatPost200Response
• SteerCompletionChatPostRequest
• SteerCompletionPost200Response
• SteerCompletionRequest
• TokenizePost200Response
• TokenizePostRequest
• UtilSaeTopkByDecoderCossimPost200Response
• UtilSaeTopkByDecoderCossimPost200ResponseTopkDecoderCossimFeaturesInner
• UtilSaeTopkByDecoderCossimPostRequest
• UtilSaeVectorPost200Response
• UtilSaeVectorPostRequest

Documentation For Authorization

Authentication schemes defined for the API:

SimpleSecretAuth

• Type: API key
• API key parameter name: X-SECRET-KEY
• Location: HTTP header

Author

[e-mail]