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

BER

-0.75

Ply

-0.72

ighed

-0.71

Nas

-0.70

bane

-0.66

Constructed

-0.64

Bron

-0.63

hirt

-0.62

ellect

-0.62

quote

-0.61

POSITIVE LOGITS

artifacts

0.77

axter

0.76

 samurai

0.74

URA

0.70

 torches

0.70

 flashback

0.69

olla

0.68

 Dino

0.68

 squid

0.66

 Armageddon

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.