INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

itÃ©

-0.74

tails

-0.71

 Druid

-0.69

izo

-0.68

rouse

-0.67

INGTON

-0.66

irlf

-0.65

aucus

-0.64

ixture

-0.64

mentation

-0.62

POSITIVE LOGITS

ãĤ´ãĥ³

0.86

apt

0.79

etheless

0.77

éĸ

0.71

ãĥīãĥ©

0.71

clips

0.68

PokÃ©mon

0.67

rha

0.64

NetMessage

0.63

 snipp

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.