wang2226/steering-showcase

Send depth, token limit and contrast continuations; keep live takeaways neutral

Reject the last decoder block as a steering layer

Report the GPU it ran on, and take the softmax on the CPU in float64

Record the CPU/GPU comparison result in the module docstrings

Read probabilities from the full softmax, and report the revision

Serve several models, and a visitor's own prompt

Cache deterministic results so a visitor costs no GPU quota

Deploy steering showcase API

initial commit