Skip to content

Fully automatic censorship removal for language models

Add educated guesses for parameter values to get the optimizer started

Philipp Emanuel Weidmann3f242369e07b
diff --git a/src/heretic/main.py b/src/heretic/main.py
index ea45b96..19bc3da 100644
--- a/src/heretic/main.py
+++ b/src/heretic/main.py
@@ -82,6 +82,10 @@ def main():
while batch_size <= settings.max_batch_size:
print(f"* Trying batch size [bold]{batch_size}[/]... ", end="")
+ # FIXME: Using the same prompt across the batch is a poor benchmark for MoE models,
+ # because it means that the same experts are active for all prompts at each
+ # token position (since we use deterministic decoding), which is substantially
+ # faster than if different experts must be accessed for each prompt.
prompts = [settings.test_prompt] * batch_size
try:
@@ -141,7 +145,7 @@ def main():
)
min_weight = trial.suggest_float("min_weight", 0, max_weight)
min_weight_distance = trial.suggest_float(
- "min_weight_distance", 0, len(model.model.model.layers) - 1
+ "min_weight_distance", 1, len(model.model.model.layers) - 1
)
print()
@@ -173,6 +177,26 @@ def main():
return -score
study = optuna.create_study()
+
+ # Educated guesses for parameter values to get the optimizer started.
+ for max_weight, max_weight_position, min_weight, min_weight_distance in [
+ (0.0, 0.0, 0.0, 0.5),
+ (1.0, 0.5, 0.0, 0.25),
+ (0.8, 0.7, 0.3, 0.4),
+ (0.9, 0.3, 0.1, 0.1),
+ (1.0, 1.0, 1.0, 1.0),
+ ]:
+ study.enqueue_trial(
+ {
+ "max_weight": max_weight,
+ "max_weight_position": max_weight_position
+ * (len(model.model.model.layers) - 1),
+ "min_weight": min_weight,
+ "min_weight_distance": min_weight_distance
+ * (len(model.model.model.layers) - 1),
+ }
+ )
+
study.optimize(objective, n_trials=settings.n_trials)
print()