INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

NESS

-0.70

peat

-0.69

GV

-0.62

vag

-0.60

 suggestion

-0.60

RED

-0.58

 criticism

-0.57

otto

-0.57

 contr

-0.57

Crit

-0.57

POSITIVE LOGITS

abama

0.73

 dexter

0.71

merce

0.65

byn

0.65

 Aviv

0.64

SourceFile

0.64

omen

0.64

 divid

0.64

âĹ¼

0.63

aceutical

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.