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

ãĥ¼ãĥĨ

-0.83

oneliness

-0.76

 Duck

-0.75

erno

-0.74

AVG

-0.71

 Stub

-0.71

 Opera

-0.70

 "$:/

-0.70

 Lumpur

-0.68

uterte

-0.68

POSITIVE LOGITS

eh

0.69

sha

0.67

andel

0.67

esis

0.63

snap

0.62

umbn

0.61

venge

0.60

ech

0.60

eal

0.60

isner

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.