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

 therap

-0.75

ICLE

-0.75

assis

-0.73

ULAR

-0.71

achev

-0.70

 imagination

-0.70

 handwriting

-0.69

redo

-0.68

antine

-0.68

audi

-0.67

POSITIVE LOGITS

uristic

0.72

nces

0.72

Reg

0.70

 approx

0.65

ights

0.64

ãĤ£

0.62

 patches

0.61

oji

0.60

gel

0.60

lessly

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.