Separates three user-visible actions that affect different state domains and must not be treated as synonyms.
Implementation evidence: this topic is grounded in the reviewed GGUF.MiRust.com source snapshot. It documents observed code and artifacts without claiming broad deployment, model quality, or production readiness.
Reset context
reset_context clears runtime token vectors, generated tokens, the visible KV-cache length, and generation diagnostics. It retains the model, tokenizer, scratch buffers, logits, sampling configuration, and browser transcript.
Clear transcript
The browser Clear control removes only JavaScript state.messages and DOM list items. It does not call the runtime and does not alter KV or token state.
Free model
free_model releases model, KV cache, scratch, and logits; resets generation diagnostics; marks the model absent; and changes the active quantization label to none.
Design rule
Future persistence and session work must name these state planes explicitly: presentation history, generation context, and model residency.
Scope
This starter page defines the questions, boundaries, evidence, and failure modes that should be recorded before a capability is presented as supported.
Engineering considerations
- Identify the source, version, target environment, and owner.
- Separate observed values from estimates and externally reported values.
- Record trade-offs, unsupported cases, and fallback behavior.
- Link performance statements to a compatible benchmark methodology.
Verification questions
- What exact artifact, revision, backend, and environment were reviewed?
- Which assumptions could change the result?
- Which data should be retained so another engineer can reproduce the conclusion?