fix(e2e): configure aibrix profile to use correct model for PII tests

The aibrix profile deploys vllm-llama3-8b-instruct as its LLM model,
but the PII detection test was hardcoded to use "general-expert" which
doesn't exist in aibrix deployments.

This commit configures the aibrix profile to set E2E_TEST_MODEL environment
variable to "vllm-llama3-8b-instruct" during setup, ensuring PII tests use
the correct model name for aibrix deployments.

Related: fix(e2e): make PII test model name configurable via environment variable

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: Yossi Ovadia <yovadia@redhat.com>

Author: yossiovadia

e2e/profiles/aibrix/profile.go

@@ -81,6 +81,9 @@ func (p *Profile) Setup(ctx context.Context, opts *framework.SetupOptions) error
 	p.verbose = opts.Verbose
 	p.log("Setting up AIBrix test environment")
 
+	// Configure PII test to use the correct model for aibrix
+	os.Setenv("E2E_TEST_MODEL", deploymentDemoLLM)
+
 	deployer := helm.NewDeployer(opts.KubeConfig, opts.Verbose)
 
 	// Track what we've deployed for cleanup on error