INDEX

Explanations

references to deception and hidden identities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

existence was himselfwall differences

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kembangan

-0.40

 africaine

-0.39

 surla

-0.39

AnimationsModule

-0.38

 Gemeinschaft

-0.37

Życiorys

-0.36

Activation

-0.35

consulté

-0.35

 spalle

-0.34

 industriels

-0.34

POSITIVE LOGITS

Personendaten

0.53

новништво

0.50

➯

0.48

 चीज़ों

0.48

 kaarangay

0.47

 PeEnEo

0.46

хьтан

0.42

CodedInputStream

0.41

Geplaatst

0.41

AccessorTable

0.40

Activations Density 0.052%

references to deception and hidden identities

existence was himselfwall differences

No Comments

No Known Activations