image: registry.stage.redhat.io/rhelai1/bootc-aws-nvidia-rhel9:1.4.4-1744735778
AWS Instance: g6.48xlarge
ilab, version 0.23.5
granite-3.1-8b-lab-v1

Steps to reproduce:
1. Create a aws instance with ami id ami-05f6fda386d6e6c73.
2.  use podman login registry.stage.redhat.io
3. Execute "sudo bootc switch registry.stage.redhat.io/rhelai1/bootc-aws-nvidia-rhel9:1.4.4-1744735778"
4. Execute "ilab config init"
5. Download models
ilab model download --repository docker://registry.stage.redhat.io/rhelai1/skills-adapter-v3 --release 1.4
ilab model download --repository docker://registry.stage.redhat.io/rhelai1/knowledge-adapter-v3 --release 1.4
ilab model download --repository docker://registry.stage.redhat.io/rhelai1/granite-3.1-8b-lab-v1 --release 1.4
ilab model download --repository docker://registry.stage.redhat.io/rhelai1/granite-3.1-8b-starter-v1 --release 1.4
ilab model download --repository docker://registry.stage.redhat.io/rhelai1/mixtral-8x7b-instruct-v0-1 --release 1.4
ilab model download --repository docker://registry.stage.redhat.io/rhelai1/prometheus-8x7b-v2-0 --release 1.4

6. Execute 
ilab model serve # Ctrl + C after gunicorn startsilab model chat
7. Use the prompt added from the file chat-cli.history or use nstructlab/chatlogs/chat_2025-04-16T17_41_27.log
As its a multiple prompts lead to hallucination.