Skip to content

Bump NNSight Version to 0.6.1 #85

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hannamw merged 1 commit into from
Feb 28, 2026
Merged

Bump NNSight Version to 0.6.1 #85

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
[project]
name = "circuit-tracer"
version = "0.4.0"
version = "0.4.1"
description = "Library for circuit tracing in language models"
readme = "README.md"
requires-python = ">=3.10"
Expand All @@ -9,7 +9,7 @@ dependencies = [
"huggingface_hub<1.0.0",
"ipykernel>=6.29.5,<7.0.0",
"ipywidgets>=8.1.7",
"nnsight>=0.5.13",
"nnsight==0.6.1",
"numpy>=1.24.0",
"pydantic>=2.0.0",
"safetensors>=0.5.0",
Expand Down
2 changes: 1 addition & 1 deletion tests/test_attribution_targets.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -758,7 +758,7 @@ def _run_custom_target_correctness(
n_samples: int = 20,
act_atol: float = 5e-4,
act_rtol: float = 1e-5,
logit_atol: float = 1e-4,
logit_atol: float = 2e-4,
logit_rtol: float = 1e-3,
):
"""Verify custom target direction feature attribution driven interventions produce expected activation/logit changes
Expand Down
4 changes: 2 additions & 2 deletions tests/test_attributions_gemma3_nnsight.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -504,8 +504,8 @@ def test_gemma3_with_dummy_clt():
assert isinstance(model, NNSightReplacementModel)

with model.zero_softcap():
verify_token_and_error_edges(model, graph)
verify_feature_edges(model, graph)
verify_token_and_error_edges(model, graph, act_rtol=5e-4)
verify_feature_edges(model, graph, act_rtol=5e-4)


@pytest.mark.skipif(not torch.cuda.is_available(), reason="CUDA not available")
Expand Down