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

gha

-0.73

 motif

-0.65

 timetable

-0.64

 scen

-0.64

GGGGGGGG

-0.63

ometers

-0.61

YC

-0.61

 sket

-0.61

ãģ£

-0.59

ometer

-0.59

POSITIVE LOGITS

 Dame

0.86

xus

0.80

yna

0.70

 Minion

0.69

uel

0.66

 Author

0.65

 Offline

0.64

 Definitive

0.62

 Ember

0.62

 Bret

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.