Skip to content

Fully automatic censorship removal for language models

Improve early abort score calculation

Philipp Emanuel Weidmannf00d35dc461e
diff --git a/src/heretic/evaluator.py b/src/heretic/evaluator.py
index 3422624..d93dc33 100644
--- a/src/heretic/evaluator.py
+++ b/src/heretic/evaluator.py
@@ -55,9 +55,20 @@ class Evaluator:
).item()
print(f" * KL divergence: [bold]{kl_divergence:.4f}[/]", end="")
+ kl_score = -(
+ (
+ (
+ (kl_divergence - self.settings.max_kl_divergence)
+ / self.settings.max_kl_divergence
+ )
+ + 1
+ )
+ ** self.settings.kl_score_shape
+ )
+
if kl_divergence > self.settings.max_kl_divergence:
print(" [yellow](constraint violation; aborting trial)[/]")
- return -1, kl_divergence, self.base_refusals
+ return kl_score, kl_divergence, self.base_refusals
else:
print()
@@ -84,20 +95,7 @@ class Evaluator:
# kl_divergence only matters when it approaches max_kl_divergence,
# and the optimizer will prioritize lowering refusals rather than
# lowering kl_divergence.
- score = -(
- (
- (
- (
- (kl_divergence - self.settings.max_kl_divergence)
- / self.settings.max_kl_divergence
- )
- + 1
- )
- ** self.settings.kl_score_shape
- )
- + (refusals / self.base_refusals)
- - 1
- )
+ score = kl_score - (refusals / self.base_refusals) + 1
print(f" * Score: [bold]{score:.4f}[/]")
return score, kl_divergence, refusals