INDEX

Explanations

varying levels of difficulty and challenges in tasks or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 13-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_13/width_65k/average_l0_75

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.13.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

 surla

-0.71

 <<<<<<<<<<<<<<

-0.69

__(/*!

-0.68

ollectionView

-0.68

Geplaatst

-0.62

mobileqq

-0.60

fml

-0.59

 Италијани

-0.59

милия

-0.57

Tikang

-0.56

POSITIVE LOGITS

 difficult

1.24

difficult

1.15

 Difficult

1.13

Difficult

1.05

 difficulty

0.96

 harder

0.96

 schwierig

0.94

 challenging

0.94

 impossible

0.93

 schwier

0.91

Activations Density 0.087%

varying levels of difficulty and challenges in tasks or concepts

No Comments

No Known Activations