INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

externalActionCode

-0.88

Mus

-0.76

ndra

-0.76

Artist

-0.75

sbm

-0.72

DAQ

-0.72

ACY

-0.71

Zam

-0.71

Baz

-0.70

ï¸

-0.70

POSITIVE LOGITS

 steril

0.72

bidden

0.71

 donated

0.68

cens

0.65

 yeast

0.63

ctrl

0.63

 incorrectly

0.63

central

0.62

 sinners

0.62

 Forbidden

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.