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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aints

-0.75

iop

-0.75

interstitial

-0.73

vind

-0.71

ucks

-0.68

 Prize

-0.65

 Harding

-0.64

osate

-0.64

 cleaners

-0.62

ublic

-0.62

POSITIVE LOGITS

 coordinate

0.71

ãĥīãĥ©

0.70

ãĥ³ãĤ¸

0.68

 Tyrann

0.68

 trig

0.68

ãĤ´ãĥ³

0.66

è¦

0.65

bones

0.63

 angle

0.61

TABLE

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.