INDEX

Explanations

references to a central location or concept

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 center

-1.01

 centre

-1.01

centre

-0.79

Center

-0.79

center

-0.75

 centres

-0.75

 centers

-0.75

 centru

-0.74

Centre

-0.74

 centro

-0.73

POSITIVE LOGITS

lgari

0.47

 initState

0.40

 MenuView

0.40

 defStyleAttr

0.37

setcounter

0.36

Inflater

0.36

 springfox

0.36

insic

0.36

 Vectors

0.36

 consum

0.35

Activations Density 0.007%

references to a central location or concept

No Comments

No Known Activations