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

 millenn

-0.78

Reviewer

-0.76

ItemImage

-0.73

 guerrilla

-0.71

Accessory

-0.70

Magikarp

-0.70

 instinct

-0.70

iosyncr

-0.69

ierrez

-0.69

 practition

-0.68

POSITIVE LOGITS

 currently

1.01

 Currently

0.85

currently

0.74

Nex

0.73

CCP

0.73

byn

0.71

ourcing

0.70

othe

0.67

 presently

0.67

edia

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.