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

bilt

-0.75

aska

-0.73

ĺħ

-0.73

BuyableInstoreAndOnline

-0.72

taboola

-0.72

 Quincy

-0.72

Ĭ

-0.71

@#&

-0.69

cv

-0.68

rency

-0.68

POSITIVE LOGITS

APR

0.76

 envy

0.75

hap

0.69

 stalk

0.67

hai

0.63

blood

0.62

blooded

0.61

 Barbar

0.61

HA

0.61

 abhor

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.