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

ADRA

-0.71

DPR

-0.68

 materials

-0.67

VD

-0.67

Osw

-0.67

byss

-0.66

pg

-0.66

PLA

-0.65

 electroly

-0.63

 odds

-0.63

POSITIVE LOGITS

1.33

0.86

din

0.76

zzi

0.75

among

0.75

sticks

0.74

iak

0.74

iuses

0.74

phant

0.74

rella

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.