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

ierrez

-0.78

utenberg

-0.73

 torque

-0.69

awatts

-0.68

 ILCS

-0.67

;;;;;;;;;;;;

-0.65

intend

-0.64

 relat

-0.64

ItemTracker

-0.64

 quote

-0.63

POSITIVE LOGITS

lish

0.73

Glass

0.70

wise

0.66

Nap

0.63

PLE

0.63

 palate

0.62

 contraceptives

0.62

Cro

0.62

PA

0.61

 cann

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.