INDEX

Explanations

references to user experience and engagement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

LEncoder

-0.59

 greateſt

-0.48

 beſt

-0.46

 tranſ

-0.46

 Conſ

-0.46

ところです

-0.46

 fuckin

-0.44

trui

-0.44

 باخ

-0.44

 Inſ

-0.43

POSITIVE LOGITS

ValueStyle

0.91

__':

0.84

 سكانية

0.80

__':

0.76

 Audiodateien

0.76

SequentialGroup

0.74

tagHelperRunner

0.71

 "..\..\..\

0.70

__":

0.69

 مشين

0.68

Activations Density 0.346%

references to user experience and engagement

No Comments

No Known Activations