INDEX

Explanations

instructions or prompts related to online activities, such as logging in, subscribing, or entering information

oai_token-act-pair · gpt-3.5-turbo

phrases related to user account and authentication processes

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

ĸļ

-0.62

uably

-0.58

ãĤ©

-0.58

acus

-0.57

ä»£

-0.56

thinkable

-0.53

 Vaugh

-0.52

Keefe

-0.51

edin

-0.51

Ò

-0.50

POSITIVE LOGITS

 ORDER

0.70

 specified

0.66

 optionally

0.60

 download

0.60

 automatically

0.60

 applicable

0.60

 default

0.59

 Cancel

0.58

 valid

0.57

 notification

0.57

Activations Density 1.873%

instructions or prompts related to online activities, such as logging in, subscribing, or entering information

phrases related to user account and authentication processes

No Comments

No Known Activations