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

 Cind

-0.73

 Celt

-0.70

links

-0.69

 stripes

-0.69

favorite

-0.68

rollers

-0.65

 WRITE

-0.65

SQ

-0.64

liners

-0.64

ioxide

-0.63

POSITIVE LOGITS

umbn

0.78

 Penguin

0.75

 Norn

0.74

 Archdemon

0.72

ratulations

0.70

thal

0.69

avin

0.64

apixel

0.62

orio

0.61

 safegu

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.