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

ulk

-0.62

ility

-0.61

clusion

-0.60

 immunity

-0.59

 hearts

-0.59

wip

-0.59

 menstru

-0.58

 radi

-0.57

inn

-0.57

POSITIVE LOGITS

kefeller

0.74

ople

0.72

 Tycoon

0.69

ramid

0.67

asca

0.64

coni

0.63

erity

0.63

Gaw

0.61

andel

0.61

 Thumbnails

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.