INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

neau

-0.74

 Tradable

-0.72

iors

-0.68

vard

-0.67

 srfAttach

-0.62

apolis

-0.61

sylv

-0.61

Asset

-0.61

respect

-0.61

ICO

-0.61

POSITIVE LOGITS

onel

0.69

iery

0.68

 Picks

0.66

 Chicken

0.66

 circumference

0.66

 scramble

0.61

cision

0.59

 Bean

0.59

oned

0.59

ZIP

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.