INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

FK

-0.79

advant

-0.77

dain

-0.74

pport

-0.74

ilateral

-0.73

ppe

-0.73

illusion

-0.71

edo

-0.71

Page

-0.71

bett

-0.70

POSITIVE LOGITS

 loud

0.72

 lantern

0.72

 precedence

0.71

 Lazarus

0.70

 Ezekiel

0.68

mos

0.67

 teleport

0.66

 decre

0.66

ILY

0.65

pel

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.