INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Somali

-0.68

arak

-0.65

 level

-0.65

Md

-0.63

icable

-0.63

eki

-0.63

agara

-0.63

theless

-0.62

MFT

-0.62

alg

-0.61

POSITIVE LOGITS

strings

0.89

 coupons

0.79

 sequ

0.75

 balloons

0.73

know

0.72

ãĤ¡

0.72

dayName

0.70

 bullpen

0.69

 coupon

0.66

çīĪ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.