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

zzle

-0.69

imei

-0.65

animate

-0.64

za

-0.63

ueller

-0.63

allery

-0.63

patrick

-0.62

 Noon

-0.61

renheit

-0.61

oscope

-0.61

POSITIVE LOGITS

LGBT

0.70

successful

0.69

hots

0.68

ANC

0.66

SPONSORED

0.65

alg

0.65

FTWARE

0.64

hot

0.64

udeb

0.63

%]

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.