INDEX

Explanations

conditional statements or instructions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 betweenstory

-0.59

bury

-0.46

 '\\;'

-0.46

AntiForgeryToken

-0.46

 antif

-0.46

DropTable

-0.45

AnimationsModule

-0.45

 Normdatei

-0.44

Geografi

-0.44

ineno

-0.44

POSITIVE LOGITS

 ComVisible

0.58

 kasarigan

0.55

 Roskov

0.39

:][

0.39

TabIndex

0.38

 pouce

0.35

 defaultstate

0.34

 productivo

0.33

 tránsito

0.33

 idéia

0.33

Activations Density 0.000%

conditional statements or instructions

No Comments

No Known Activations