remove max tokens from experiment

The experiment asks a very simple question so the default LLM output
won't be huge. But there's a risk of users altering the question
without noticing the token limit

Author: realark

examples/src/main/java/dev/braintrust/examples/OpenAIInstrumentationExample.java

@@ -28,7 +28,6 @@ public static void main(String[] args) throws Exception {
                             .model(ChatModel.GPT_4O_MINI)
                             .addSystemMessage("You are a helpful assistant")
                             .addUserMessage("What is the capital of France?")
-                            .maxTokens(50L)
                             .temperature(0.0)
                             .build();
             var response = openAIClient.chat().completions().create(request);