INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icle

-0.75

isman

-0.74

eed

-0.73

ause

-0.73

eding

-0.72

NESS

-0.72

OWN

-0.71

ONS

-0.71

itled

-0.68

icles

-0.68

POSITIVE LOGITS

 snipp

0.70

 Tour

0.70

mercial

0.67

tesy

0.67

 lifes

0.65

emort

0.65

paren

0.64

 fortun

0.63

 whis

0.62

 souven

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.