FA-98417 / Token sampling filters / Member archive

Token sampling: Adaptive top-a uses an absolute rather than relative gap · case 02

Adaptive top-a uses an absolute rather than relative gap.

Member previewVariant 2 · 3 implementations · 6 checks per implementation

Case contract

Sort descending with lexical ties; retain the best token and following tokens while their logit is within a times the best-to-second-best gap. a<=0 retains only the best; singleton and empty inputs are handled.

Why this case matters

A finite, deterministic token-selection contract isolates one decoding operation from model inference and runtime behavior.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixture	Actual	Expected	Outcome
scaled scores	`["a"]`	`["a", "b", "c"]`	Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

Member access is invitation-based. Sign in with your invited account to inspect the sources.