INDEX

Explanations

references to the self, particularly in the context of personal experiences and perspectives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

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

shadowOffset

-0.38

the

-0.35

 salvaje

-0.33

 Abonnez

-0.33

bibfield

-0.32

 Bhattach

-0.32

 clássico

-0.31

 nocturno

-0.31

 húmedo

-0.30

 uden

-0.30

POSITIVE LOGITS

me

0.93

Me

0.88

 Myself

0.82

Me

0.81

myself

0.81

 estekak

0.77

rungsseite

0.77

saraba

0.76

SuspendLayout

0.76

DockStyle

0.74

Activations Density 0.025%

references to the self, particularly in the context of personal experiences and perspectives

No Comments

No Known Activations