INDEX

Explanations

phrases indicating instructions or guidance for technical processes and settings

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

 CanadaChoose

-0.46

']>

-0.38

bali

-0.37

))/

-0.37

ViewInit

-0.35

нове

-0.35

numerusform

-0.35

Val

-0.34

↑

-0.34

€

-0.34

POSITIVE LOGITS

 للاسماء

0.68

 wikipagina

0.62

WebVitals

0.61

 חיצוניים

0.59

oa̍t

0.57

ContentAsync

0.56

 hObject

0.56

 propOrder

0.55

 يتيمه

0.54

medriver

0.52

Activations Density 0.190%

phrases indicating instructions or guidance for technical processes and settings

No Comments

No Known Activations