INDEX

Explanations

locations and geographic references

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

providedIn

-0.71

Ӕ

-0.66

 gynhyrchwyd

-0.65

 Italijanski

-0.60

 ujednoznacz

-0.59

 kasarigan

-0.59

titleMargin

-0.59

featureID

-0.58

portál

-0.58

Personendaten

-0.57

POSITIVE LOGITS

Visitor

0.45

ắn

0.34

 area

0.33

kyard

0.33

Zaj

0.31

uxxxx

0.31

upol

0.31

 visitors

0.30

 Visitor

0.30

 Visitors

0.30

Activations Density 0.245%

locations and geographic references

No Comments

No Known Activations