INDEX

Explanations

emergency

np_max-act · gemini-2.0-flash

This neuron detects tokens related to providing or requesting contact and emergency-contact information (e.g., “reach,” “phone,” “emergency,” “contact”).

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

 lamin

-0.07

�

-0.07

 Near

-0.07

	string

-0.07

 morale

-0.06

LOSE

-0.06

 Archie

-0.06

	lines

-0.06

$key

-0.06

 BOTH

-0.06

POSITIVE LOGITS

ilians

0.06

 reactors

0.06

 reactor

0.06

 quyết

0.06

čer

0.06

 Monk

0.06

�

0.06

_supplier

0.06

 cháy

0.06

.inc

0.06

Activations Density 0.007%

emergency

This neuron detects tokens related to providing or requesting contact and emergency-contact information (e.g., “reach,” “phone,” “emergency,” “contact”).

No Comments

No Known Activations