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

lly

-0.79

xual

-0.78

ĸļ

-0.73

 rall

-0.70

-0.68

-0.66

 experien

-0.66

ayne

-0.65

olds

-0.64

POSITIVE LOGITS

inance

0.81

atos

0.78

hots

0.74

HS

0.71

dust

0.69

acht

0.69

GW

0.67

INC

0.66

cer

0.65

aic

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.