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

 propensity

-0.68

 props

-0.67

 explan

-0.66

yrim

-0.65

awaru

-0.65

 goodies

-0.64

 theme

-0.62

ok

-0.62

 pockets

-0.60

 dism

-0.60

POSITIVE LOGITS

ciplinary

0.92

ibel

0.89

peak

0.75

hare

0.73

ibe

0.72

lled

0.67

Test

0.66

sites

0.65

acter

0.65

quit

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.