INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

jri

-0.82

ilver

-0.81

ellery

-0.77

inyl

-0.76

bags

-0.76

itÃ©

-0.76

liga

-0.75

lus

-0.73

dinand

-0.72

ibli

-0.71

POSITIVE LOGITS

 Atmosp

0.66

 Plain

0.66

 Vari

0.63

FUL

0.61

 Anchorage

0.61

 Example

0.60

 Variant

0.59

 Personally

0.59

 shorth

0.58

 Limit

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.