INDEX

Explanations

names and identifiers of geographic locations and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 løpet

-0.38

Traceback

-0.36

 lägen

-0.34

 klubben

-0.33

 nästa

-0.33

 detalj

-0.32

 syke

-0.32

 skolen

-0.31

 eneste

-0.31

ApiModelProperty

-0.30

POSITIVE LOGITS

WriteTagHelper

0.71

BeginContext

0.70

liceerd

0.61

OGND

0.60

 مرئيه

0.59

 surla

0.59

LabelTagHelper

0.58

$_['

0.57

ViewImports

0.57

 ***!

0.57

Activations Density 0.362%

names and identifiers of geographic locations and entities

No Comments

No Known Activations