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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥ©ãĥ³

-0.61

QC

-0.61

 Copenhagen

-0.59

 Cors

-0.59

 segments

-0.58

ãĥ

-0.58

 allerg

-0.58

bf

-0.57

 Header

-0.57

ilde

-0.57

POSITIVE LOGITS

yton

0.80

gres

0.73

thood

0.67

oun

0.65

creen

0.65

psy

0.64

acebook

0.64

ikarp

0.64

miah

0.62

LINE

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.