INDEX

Explanations

names and terminology related to scientific procedures and analysis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

ed

-0.85

DispatchToProps

-0.78

 Helium

-0.73

 Carrington

-0.70

moveToFirst

-0.69

 Moha

-0.68

istani

-0.68

 Bragg

-0.67

ように

-0.67

 HAST

-0.66

POSITIVE LOGITS

彡

0.71

er

0.67

IRECT

0.67

 Dougherty

0.66

perties

0.64

Evans

0.64

zeria

0.64

erii

0.64

 ExecuteAsync

0.62

odendron

0.62

Activations Density 0.609%

names and terminology related to scientific procedures and analysis

No Comments

No Known Activations