INDEX

Explanations

phrases related to confidentiality and private information

oai_token-act-pair · gpt-3.5-turbo

mentions of confidentiality and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

annis

-0.95

ð

-0.79

Ì¶

-0.78

 Archdemon

-0.71

iph

-0.70

uph

-0.70

apple

-0.68

akings

-0.68

ixels

-0.68

udeb

-0.67

POSITIVE LOGITS

 confidential

1.44

 informant

1.20

 informants

1.10

idential

1.08

 confidentiality

1.03

 correspondence

0.86

 documents

0.84

 briefings

0.83

 transmissions

0.81

 memos

0.81

Activations Density 0.005%

phrases related to confidentiality and private information

mentions of confidentiality and related terms

No Comments

No Known Activations