INDEX

Explanations

phrases related to planning or decision-making

oai_token-act-pair · gpt-3.5-turbo

terms associated with crime and illicit activities

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

Lup

-0.75

 Rooney

-0.74

Lep

-0.70

 Corona

-0.69

 Wast

-0.65

Klu

-0.63

Sau

-0.63

Ku

-0.63

utenberg

-0.62

 Roof

-0.62

POSITIVE LOGITS

rieg

0.81

éĸ

0.77

 moderation

0.75

ï¸

0.75

é¾įå¥ĳå£«

0.74

jamin

0.73

OSED

0.73

imester

0.70

istan

0.69

Introduced

0.68

Activations Density 0.131%

phrases related to planning or decision-making

terms associated with crime and illicit activities

No Comments

No Known Activations