September 19th 2024

We are delighted to introduce our latest features from the recent period, enhancing your experience with improved functionality and performance.

Tools Support in Session Explorer

Gain insights into the detailed usage of tools within each user-LLM session using the enhanced Session Explorer. Key updates include:

  1. Overview Tab: A chat-like interface displaying the full session, including tool requests and responses.
  2. Tools Tab: Lists all tools used during the session, including their names, descriptions, and parameters.
  3. Extractions Tab: Shows content extracted from the session.
  4. Metadata Tab: Demonstrates all the policies that were enabled during the session, highlights the triggered policies (which detected violations), and the action taken by Aporia. The tab also displays total token usage, estimated session cost, and the LLM model used.

These updates provide full visibility into all aspects of user-LLM interactions.

New PII Category: Location

We have expanded PII detection capabilities with the addition of the location category, which identifies geographical details in sensitive data, such as ‘West End’ or ‘Brookfield.’

Dataset Upload

We’re excited to introduce the Dataset Upload feature, enabling you to upload datasets directly to Aporia for review and analysis. Supported file format is CSV (max 20MB), with at least one filled column for ‘Prompt’ or ‘Response‘.

September 13, 2024

Prompt Injection Fine-Tuned Policy

We’ve refined our prompt injection policy to enhance performance with three sensitivity levels. This new approach allows you to select the sensitivity level that best suits your use case. The levels are defined as:

  1. Level 1: Detects only clear cases of prompt injection. Ideal for minimizing false positives but might overlook ambiguous cases.
  2. Level 2: Balanced detection. Effectively identifies clear prompt injections while reasonably handling ambiguous cases.
  3. Level 3: Detects most prompt injections, including ambiguous ones.

PII Masking – New PII Policy Action

We’ve introduced a new action for our PII policy; PII masking, that replaces sensitive data with corresponding tags before the message is processed or sent. This ensures that sensitive information remains protected while allowing conversations to continue.

Example Before Masking:

Please send the report to john.doe@example.com and call me at 123-456-7890.

Example After Masking:

Please send the report to <EMAIL> and call me at <PHONE_NUMBER>.

API Keys Management

We’ve added a new API Keys table under the “My Account” section to give you better control over your API keys. You can now create and revoke API keys. For security reasons, you won’t be able to view the key again after creation, so if you lose this secret key, you’ll need to create a new one.

Navigation Between Dashboard and Projects

General Dashboard: You can now easily navigate from the general dashboard to your projects by simply clicking on any project in the active project section.

Project Dashboard: Clicking on any action or policy will take you directly to the project’s Session Explorer, pre-filtered by the same policy/action and date range. Additionally, “Clicking on the prompt/response graphs in the analytics report will also navigate you to the Session Explorer, filtered by the corresponding date range.

Policy Example Demonstrations

We’ve enhanced the examples section for each policy to provide clearer explanations. You can now view a sample conversation between a user and an LLM when a violation is detected and action is taken by Aporia. Simply click on “Examples” before adding a policy to your project to see which violations each policy is designed to prevent.

Improved Policy Configuration Editing

We’ve streamlined the process of editing custom policy configurations. Now, when you click “Edit Configuration”, you’ll be taken directly to the policy configuration page in the policy catalog. Once there, you can easily return to your project with the new “Back to Project” arrow.

August 20, 2024

We are delighted to introduce our latest features and fixes from the recent period, enhancing your experience with improved functionality and performance.

New Dashboards

We have developed new dashboards that allow you to view both a general organizational overview and specific project-focused insights. View total messages and detected prompts and responses violations over time with enhanced filtering and sorting options. See which policies triggered violations and the actions taken by Aporia’s Guardrails.


Restricted Phrases Policy

We have implemented the Restricted Phrases Policy to manage compliance by preventing the use of specific terms or phrases in LLM responses. This policy identifies and handles prohibited language, ensuring that any flagged content is either logged, overridden, or rephrased to maintain compliance.

restricted phrases

 

Navigate Between Aporia’s Platforms

We have made it easier for you to navigate between Aporia’s Gen AI Platform and Classic ML Platform. A new icon at the top of the site allows seamless switching between the two platforms.

aporia platform navigation

Policy Threat Level

We have introduced a new feature that allows you to assign a threat level to each policy, indicating its criticality (Low, Substantial, Critical). This setting is displayed across your dashboards, helping you manage prompts and responses violations effectively.

threat level

Policy Catalog Search Bar

We have added a search bar to the policy catalog, allowing you to perform context-sensitive searches.

policy catalog search bar

August 6, 2024

We are delighted to introduce our latest features and fixes from the recent period, enhancing your experience with improved functionality and performance.

Task Adherence Policy

We have introduced a new policy to ensure that user messages and assistant responses strictly adhere to the tasks and objectives outlined in the policy. This policy evaluates each prompt or response to ensure alignment with the conversation’s goals.

Language Mismatch Policy

We have created a new policy that detects when the LLM responds to a user’s question in a different language. The policy allows you to choose a new action, “Translate response” which will translate the response to the user’s prompt language.

Integration page

We are happy to introduce our new Integration page! Easily connect your LLM applications through AI Gateways integrated with Aporia, Aporia’s REST API and OpenAI Proxy, with detailed guides and seamless integration options.

Project Cards

We have updated the project overview page to provide more relevant information at a glance. Each project now displays its name, icon, size, integration status, description, and active policies. Quick actions such as integrating your project and activating policies, are available to enhance your experience.