INDEX

Explanations

apologize

np_max-act · gemini-2.0-flash

This neuron detects tokens expressing an apology (e.g., “apologise,” “apologized,” etc.).

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

 resistor

-0.07

-learning

-0.07

[{

-0.06

 highlights

-0.06

 foundation

-0.06

 myList

-0.06

	padding

-0.06

 interactions

-0.06

 конс

-0.06

 Interactive

-0.06

POSITIVE LOGITS

 apologized

0.10

 apologise

0.09

 apologies

0.08

 apology

0.07

ابعة

0.07

peq

0.07

ÃO

0.07

 apologize

0.07

likely

0.07

PAY

0.07

Activations Density 0.007%

apologize

This neuron detects tokens expressing an apology (e.g., “apologise,” “apologized,” etc.).

No Comments

No Known Activations