INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 divest

-0.85

owe

-0.77

 affirmative

-0.72

hemy

-0.70

 undes

-0.68

ilion

-0.65

 Owens

-0.63

 Americ

-0.63

 assert

-0.63

oulos

-0.62

POSITIVE LOGITS

£ı

0.73

mask

0.72

 Mandal

0.70

 pubs

0.69

rontal

0.68

aldi

0.68

Sto

0.66

MY

0.65

agar

0.65

eral

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.