choice logits

Hold

Techniques

Model output scores used to represent preference over options.

Why it's here

Placed in Hold: 1 article(s) of evidence from 1 source(s), led by research-stage coverage, with 1 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.

Evidence (1)

2Hacker News·9/18/2026research
OpenJev direct readout for choice probabilities
The post describes a direct readout method for choice probabilities by taking a model’s choice logits and normalizing only across the options provided, without additional decoding. It appears to be a concise technical note rather than a full product or release announcement.