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

apo

-0.74

alia

-0.69

clus

-0.69

andi

-0.67

osition

-0.67

notations

-0.67

oan

-0.65

jug

-0.63

 Button

-0.63

 Explicit

-0.62

POSITIVE LOGITS

çīĪ

0.78

inson

0.71

 Siber

0.70

 Planes

0.70

obin

0.68

pts

0.64

Fans

0.62

enegger

0.62

ä»£

0.60

Guest

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.