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

worth

-0.72

vest

-0.69

ket

-0.69

achy

-0.69

oute

-0.69

ards

-0.68

mg

-0.68

Picture

-0.67

llo

-0.66

ames

-0.65

POSITIVE LOGITS

displayText

0.79

 Intake

0.68

aundering

0.66

 Chop

0.65

 Gamb

0.64

EStreamFrame

0.64

 Shant

0.64

 Kumar

0.64

 documentary

0.62

é¾

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.