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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agall

-0.80

etsk

-0.74

alid

-0.71

ãĤ¦ãĤ¹

-0.65

ernandez

-0.64

REM

-0.64

Alb

-0.64

ãĤ¶

-0.64

andum

-0.63

adiq

-0.63

POSITIVE LOGITS

unks

0.71

0.70

 Vendor

0.70

 ---------

0.70

yrics

0.67

 Wear

0.66

 Poster

0.66

sf

0.64

 Codec

0.62

shed

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.