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

=]

-0.77

 Halifax

-0.73

enstein

-0.71

inet

-0.70

 reservations

-0.69

thing

-0.66

 Tolkien

-0.65

":["

-0.65

NEY

-0.64

bid

-0.64

POSITIVE LOGITS

heed

0.72

Ĥİ

0.66

velength

0.65

ichen

0.64

heng

0.62

axy

0.61

 bullet

0.60

idav

0.60

creen

0.60

ogyn

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.