probekit 0.1.0

A versatile kit for training and using linear probes on neural network activations.

Probes

A lightweight, modular library for training linear probes and steering vectors on neural network activations.

Core Design (V2)

This library separates Semantics (the probe model) from Fitting (how it's learned).

1. The Models: LinearProbe and ProbeCollection

• LinearProbe (probes.core.probe): A container for a single probe (+ normalization stats).
• ProbeCollection (probes.core.collection): A container for a batch of probes.
  • to_tensor(): Stacks weights into [B, D] and biases into [B].
  • best_layer(metric): Finds the probe with the best validation accuracy.

2. The Fitters

Functional solvers in probes.fitters take training data and return a LinearProbe (or ProbeCollection).

• fit_logistic: Standard L2-regularized Logistic Regression.
• fit_elastic_net: ElasticNet (L1 + L2), useful for sparse features (SAEs, Neurons).
• fit_dim: Difference-in-Means (Class 1 Mean - Class 0 Mean).

Batched GPU Fitters

Optimized PyTorch implementations in probes.fitters.batch handle 3D inputs [B, N, D] efficiently on GPU:

• fit_logistic_batch: Batched IRLS solver.
• fit_dim_batch: Vectorized DiM with median thresholding.
• fit_elastic_net_path: Efficiently fits a regularization path (multiple alphas) using warm-starting.

Quick Start

The high-level API automatically routes based on the input dimensions:

from probes import sae_probe, dim_probe

# 1. Single Probe (X: [N, D], y: [N])
probe = sae_probe(X_2d, y_1d)

# 2. Batched Probes (X: [B, N, D], y: [B, N] or [N])
# Automatically uses GPU fitters and returns a ProbeCollection
probes = sae_probe(X_3d, y)
weights, biases = probes.to_tensor() # [B, D], [B]

Steering Vectors

You can build steering vectors for individual probes or entire collections:

from probes import build_steering_vector, build_steering_vectors

# Single
vec = build_steering_vector(probe, sae_model, layer=10)

# Batched (Maps layers to probes)
vecs = build_steering_vectors(probe_collection, sae_model, layers=[8, 9, 10])

Structure

• probes/core/: LinearProbe and ProbeCollection definitions.
• probes/fitters/:
  • logistic.py, elastic.py, dim.py: Single-probe (CPU/sklearn) fitters.
  • batch/: Optimized GPU-batched fitters (IRLS, ISTA, DiM).
• probes/api.py: High-level aliases and dimension routing.
• probes/steering/: Tools for building steering vectors.