INDEX

Explanations

geographic and administrative terms related to locations

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

FieldNumber

-0.57

isis

-0.45

 /*#__

-0.44

rader

-0.44

 gründ

-0.42

="@+

-0.42

="&#

-0.42

yak

-0.42

 endregion

-0.41

̈́

-0.40

POSITIVE LOGITS

 pinulongan

0.52

 AssemblyCompany

0.52

 transfieras

0.42



0.41

CppMethod

0.41

 فريبيس

0.40

 politiques

0.39

 almofada

0.39

Gemeinsame

0.39

Politique

0.38

Activations Density 0.006%

geographic and administrative terms related to locations

No Comments

No Known Activations