INDEX

Explanations

truly

np_max-act · gemini-2.0-flash

This neuron fires on emphatic praise phrases—particularly the “truly a …”‐style intensifier pattern used for glowing descriptions.

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

 lick

-0.08

ивання

-0.08

.simpleButton

-0.07

 Blackburn

-0.07

 ()=>

-0.07

宮

-0.07

اعية

-0.07

 anyway

-0.06

 Snyder

-0.06

 Nack

-0.06

POSITIVE LOGITS

 truly

0.16

 Truly

0.10

Taj

0.07

 resembl

0.07

 really

0.07

 fully

0.07

 genuinely

0.06

 amenities

0.06

 ویر

0.06

 NSObject

0.06

Activations Density 0.013%

truly

This neuron fires on emphatic praise phrases—particularly the “truly a …”‐style intensifier pattern used for glowing descriptions.

No Comments

No Known Activations