INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

chev

-0.66

osphere

-0.63

nov

-0.62

 obser

-0.62

selves

-0.62

 Appeals

-0.60

 Kavanaugh

-0.58

stasy

-0.57

asure

-0.57

gemony

-0.57

POSITIVE LOGITS

erate

0.70

 Dill

0.66

Redd

0.65

»Ĵ

0.65

ezvous

0.65

onite

0.63

Kid

0.62

IER

0.62

idth

0.60

 Hallow

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.