La Plateforme

¶Moderation

Chat moderation with mistral is a classifier model based on Ministral. Mistral released two endpoints: https://api.mistral.ai/v1/chat/moderations and https://api.mistral.ai/v1/moderations

Get the full response from la plateforme for your input message for moderation. No moderation is needed for this message.

Tip

with the filter parameter set to true, the moderation() method will output only the moderated items

$client = new MistralClient($apiKey);

try {
    $result = $client->moderation(model: 'mistral-moderation-latest' , input: 'you are a very nice person', filter:false);
} catch (\Throwable $e) {
    echo $e->getMessage();
    exit(1);
}

Result

 Array
(
    [id] => 869e8cdec2d74e3ba5ecd2b310f75dc5
    [usage] => Array
        (
            [prompt_tokens] => 9
            [total_tokens] => 98
            [completion_tokens] => 0
            [request_count] => 1
        )

    [model] => mistral-moderation-latest
    [results] => Array
        (
            [0] => Array
                (
                    [category_scores] => Array
                        (
                            [sexual] => 1.8925149561255E-5
                            [hate_and_discrimination] => 6.2050494307186E-5
                            [violence_and_threats] => 0.00020342698553577
                            [dangerous_and_criminal_content] => 7.0311849412974E-5
                            [selfharm] => 6.1441742218449E-6
                            [health] => 1.1478768101369E-5
                            [financial] => 4.7850944611127E-6
                            [law] => 5.093705112813E-6
                            [pii] => 0.00010889690747717
                        )

                    [categories] => Array
                        (
                            [sexual] =>
                            [hate_and_discrimination] =>
                            [violence_and_threats] =>
                            [dangerous_and_criminal_content] =>
                            [selfharm] =>
                            [health] =>
                            [financial] =>
                            [law] =>
                            [pii] =>
                        )

                )

        )

)

$client = new MistralClient($apiKey);

try {
    $result = $client->moderation(model: 'mistral-moderation-latest' , input: ['you are a disgusting person'], filter:true);
} catch (\Throwable $e) {
    echo $e->getMessage();
    exit(1);
}

Result

Array
(
    [0] => Array
        (
            [0] => hate_and_discrimination
        )

)

¶Chat moderation

Chat moderation will work in a very similar way. Except that it will use a Messages object with a multi-turn conversation session.

$client = new MistralClient($apiKey);
$messages = $client->getMessages()->addUserMessage('What is the best French cheese?');

try {
    $result = $client->chatModeration(model: 'mistral-moderation-latest', messages: $messages, filter: false);
} catch (\Throwable $e) {
    echo $e->getMessage();
    exit(1);
}