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

uana

-0.96

uyomi

-0.93

iple

-0.89

ipal

-0.88

kan

-0.83

azel

-0.82

ilipp

-0.82

phies

-0.82

auga

-0.81

uese

-0.79

POSITIVE LOGITS

 Boko

0.66

 fame

0.64

Xue

0.63

 bureaucr

0.62

 Panda

0.62

 Darling

0.62

 reusable

0.61

 Tian

0.61

 Zhao

0.60

 bears

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.