INDEX

Explanations

ped

np_max-act · gemini-2.0-flash

queries related to customer service interactions and responses.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

This neuron activates on references to video-game or gaming mechanics terms (like “game,” “skill tree,” “torpedoes,” “missile launcher,” 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 relacion

-0.06

 relaxing

-0.06

ola

-0.06

丹

-0.06

ermint

-0.06

µ

-0.06

 podnik

-0.06

ILD

-0.06

 culpa

-0.06

alling

-0.06

POSITIVE LOGITS

 Alic

0.07

ended

0.07

(Path

0.06

=text

0.06

?",

0.06

Bin

0.06

↵↵

0.06

.SIZE

0.06

=[

0.06

різ

0.06

Activations Density 0.569%

ped

queries related to customer service interactions and responses.

This neuron activates on references to video-game or gaming mechanics terms (like “game,” “skill tree,” “torpedoes,” “missile launcher,” etc.).

No Comments

No Known Activations