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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Wrestle

-0.69

 Serge

-0.64

 Initiative

-0.63

 Michaels

-0.62

 Hudson

-0.62

Xan

-0.62

 Alter

-0.62

 Randall

-0.61

 Veterans

-0.61

 unrecogn

-0.60

POSITIVE LOGITS

Updated

0.88

arcity

0.85

idated

0.77

enza

0.77

chnology

0.76

merce

0.75

hetti

0.74

illin

0.74

adian

0.72

nered

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.