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

 Advertisement

-0.78

 Gork

-0.69

ctuary

-0.68

 Tong

-0.67

TPPStreamerBot

-0.65

itor

-0.65

Mal

-0.64

iannopoulos

-0.64

Scar

-0.62

 Gast

-0.62

POSITIVE LOGITS

ãĥĺ

0.83

elsen

0.82

lÃ©

0.75

agues

0.68

fine

0.68

ctrl

0.67

querque

0.67

enson

0.67

ossibility

0.67

apo

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.