ai-space/heretic
3
0
Fork 0
mirror of https://github.com/p-e-w/heretic.git synced 2026-06-01 20:58:47 +02:00
Code Issues Packages Projects Releases Wiki Activity

docs: improve settings documentation

Browse Source
This commit is contained in:
Philipp Emanuel Weidmann
2026-02-11 10:19:05 +05:30
parent 10ceb3098e
commit b873598b77
3 changed files with 12 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
config.default.toml
+6 -3
View File
@@ -1,4 +1,5 @@
# Copy this file to config.toml and edit the configuration to your liking.
# Rename this file to config.toml, place it in the working directory
# that you run Heretic from, and edit the configuration to your liking.
# List of PyTorch dtypes to try when loading model tensors.
# If loading with a dtype fails, the next dtype in the list will be tried.
@@ -77,9 +78,11 @@ row_normalization = "none"
# larger output files and may slow down evaluation.
full_normalization_lora_rank = 3
# The symmetric winsorization to apply to each layer of the per-prompt residuals,
# The symmetric winsorization to apply to the per-prompt, per-layer residual vectors,
# expressed as the quantile to clamp to (between 0 and 1). Disabled by default.
# Example: winsorization_quantile = 0.95 applies a 95% winsorization.
# This can tame so-called "massive activations" that occur in some models.
# Example: winsorization_quantile = 0.95 computes the 0.95-quantile of the absolute values
# of the components, then clamps the magnitudes of all components to that quantile.
winsorization_quantile = 1.0
# Number of abliteration trials to run during optimization.
config.noslop.toml
+2 -1
View File
@@ -1,4 +1,5 @@
# Copy this file to config.toml and edit the configuration to your liking.
# Rename this file to config.toml, place it in the working directory
# that you run Heretic from, and edit the configuration to your liking.
max_response_length = 300
src/heretic/config.py
+4 -2
View File
@@ -207,9 +207,11 @@ class Settings(BaseSettings):
winsorization_quantile: float = Field(
default=1.0,
description=(
"The symmetric winsorization to apply to each layer of the per-prompt residuals, "
"The symmetric winsorization to apply to the per-prompt, per-layer residual vectors, "
"expressed as the quantile to clamp to (between 0 and 1). Disabled by default. "
"Example: winsorization_quantile = 0.95 applies a 95% winsorization."
'This can tame so-called "massive activations" that occur in some models. '
"Example: winsorization_quantile = 0.95 computes the 0.95-quantile of the absolute values "
"of the components, then clamps the magnitudes of all components to that quantile."
),
)