INDEX

Explanations

high-frequency terms and concepts related to technical specifications and regulations

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

ⓧ

-0.66

setVerticalGroup

-0.63

verifyException

-0.60

GOTREF

-0.59

__':

-0.59

tonsoft

-0.58

__":

-0.57

 OMITBAD

-0.56

]<<"

-0.55

agisse

-0.55

POSITIVE LOGITS

VersionUID

0.38

 šť

0.36

guantes

0.35

 frutta

0.33

 estómago

0.33

czu

0.33

getOut

0.32

 mercancías

0.32

 Erkrank

0.31

 cucchiaio

0.31

Activations Density 0.043%

high-frequency terms and concepts related to technical specifications and regulations

No Comments

No Known Activations