INDEX

Explanations

phrases starting with "Do."

oai_token-act-pair · gpt-3.5-turbo

Imperative phrases or calls to action

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Reviewer

-0.80

ulates

-0.75

workshop

-0.75

ulatory

-0.66

boarding

-0.66

Gaza

-0.66

ItemTracker

-0.65

ãĤ¼ãĤ¦ãĤ¹

-0.65

tnc

-0.64

dayName

-0.64

POSITIVE LOGITS

omsday

1.24

zens

1.14

herty

1.11

ppel

1.10

lyak

0.93

lez

0.93

oley

0.91

ctr

0.86

orman

0.83

berman

0.81

Activations Density 0.027%

phrases starting with "Do."

Imperative phrases or calls to action

No Comments

No Known Activations