INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Ķ

-0.74

raq

-0.70

 Pyth

-0.69

ģĸ

-0.66

 palms

-0.64

 prox

-0.63

Ĥ¬

-0.62

²¾

-0.62

 reflex

-0.62

soever

-0.61

POSITIVE LOGITS

haven

0.85

bourg

0.76

ancial

0.75

resents

0.72

hesda

0.71

udeau

0.70

vision

0.69

Indian

0.68

Mac

0.68

stown

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.