INDEX

Explanations

terms related to official documents or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 שוליים

-0.69

enumi

-0.65

expandindo

-0.61

enumii

-0.58

writeField

-0.58

 Paix

-0.54

 surla

-0.53

 Vitae

-0.53

 урна

-0.53

ագրություններ

-0.53

POSITIVE LOGITS

OFFICIAL

0.81

dom

0.76

 OFFICIAL

0.76

 sanctioned

0.74

 unofficial

0.69

 officiels

0.69

 official

0.68

Portail

0.67

 Official

0.61

 resmi

0.60

Activations Density 0.048%

terms related to official documents or entities

No Comments

No Known Activations