logprob_result.h File Reference

Per-token log-probability evaluation result. More...

#include <cstdint>
#include <vector>

Include dependency graph for logprob_result.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
struct	entropic::LogprobResult
	Per-token log-probability evaluation result. More...

Namespaces
namespace	entropic
	Activate model on GPU (WARM → ACTIVE).

Detailed Description

Per-token log-probability evaluation result.

Returned by InferenceBackend::evaluate_logprobs(). Contains per-token log-probabilities, the input tokens echoed back for verification, and aggregate metrics (perplexity, total log-probability).

Indexing

For a sequence of N input tokens, logprobs contains N-1 values: logprobs[i] = log P(tokens[i+1] | tokens[0..i])

Perplexity

perplexity = exp(-1/(N-1) * sum(logprobs))

Version

1.9.10

Definition in file logprob_result.h.