INDEX

Explanations

terms related to scientific or technical measurements and findings

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

ंदीखरीदारी

-0.78

ſammen

-0.67

XmlAccessorType

-0.67

 zwiſchen

-0.66

protoimpl

-0.65

 verſch

-0.65

iſchen

-0.64

 ſeines

-0.64

 unſer

-0.64

ロウィン

-0.64

POSITIVE LOGITS

↵↵

0.35

eso

0.35

0.34

↵

0.33

Hentet

0.32

btnAdd

0.31

此

0.31

Transporte

0.30

<h2>

0.30

ayr

0.30

Activations Density 0.315%

terms related to scientific or technical measurements and findings

No Comments

No Known Activations