INDEX

Explanations

language related to consent to receive promotional offers

oai_token-act-pair · gpt-3.5-turbo

phrases related to promotional offers and marketing communications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

issors

-0.56

 Heritage

-0.56

 Antiqu

-0.52

zik

-0.52

FAC

-0.51

 imped

-0.51

miah

-0.50

 Arts

-0.50

Tac

-0.50

mand

-0.48

POSITIVE LOGITS

htm

0.65

Detailed

0.61

ctory

0.59

icro

0.57

olicited

0.55

padding

0.55

html

0.55

Reviewer

0.55

 dialog

0.53

0.52

Activations Density 0.018%

language related to consent to receive promotional offers

phrases related to promotional offers and marketing communications

No Comments

No Known Activations