INDEX

Explanations

Problem, solution

np_max-act · gemini-2.0-flash

The neuron detects generic methodological-introduction phrases (e.g. “one approach…,” “a common technique…,” “one of the methods…,” “the proposal for…,” “used to achieve…,” “to analyze…”) that signal the description of a method or approach.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 involves

-0.08

ैप

-0.07

 не

-0.07

усти

-0.07

 eleven

-0.07

 Neither

-0.07

County

-0.07

KIT

-0.06

ovo

-0.06

 neither

-0.06

POSITIVE LOGITS

्डल

0.07

-------------</

0.06

лас

0.06

 rhythms

0.06

 artifacts

0.06

 κύ

0.06

UniformLocation

0.06

 cracks

0.06

 Homework

0.06

 rift

0.05

Activations Density 0.064%

Problem, solution

The neuron detects generic methodological-introduction phrases (e.g. “one approach…,” “a common technique…,” “one of the methods…,” “the proposal for…,” “used to achieve…,” “to analyze…”) that signal the description of a method or approach.

No Comments

No Known Activations