INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Madness

-0.89

 Moment

-0.75

kn

-0.73

Me

-0.67

Yo

-0.67

MO

-0.66

Hod

-0.65

âĦ¢:

-0.65

Mug

-0.64

Kot

-0.64

POSITIVE LOGITS

ricane

0.76

committee

0.72

angelo

0.72

roman

0.71

uberty

0.70

arantine

0.67

eware

0.67

ranch

0.65

 Albania

0.65

 stray

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.