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

/+

-0.83

âĹ¼

-0.75

ptic

-0.74

EMOTE

-0.71

rov

-0.70

Ñģ

-0.68

PS

-0.67

bers

-0.65

 autop

-0.64

][/

-0.63

POSITIVE LOGITS

arty

0.69

 nutshell

0.66

Lab

0.66

naire

0.65

 cler

0.64

brew

0.64

KEN

0.62

stadt

0.62

 Families

0.62

 Scarlet

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.