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

oric

-0.77

Supported

-0.75

ashtra

-0.70

angelo

-0.69

Rated

-0.66

 guiIcon

-0.66

bags

-0.64

bag

-0.63

reenshots

-0.61

Bern

-0.61

POSITIVE LOGITS

 mailing

0.72

 fert

0.66

 Tues

0.65

 Tinder

0.64

aunder

0.64

Tee

0.63

bye

0.63

 fuzz

0.62

tnc

0.61

 Tasman

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.