INDEX

Explanations

phrases related to being out of a certain area or off-duty

oai_token-act-pair · gpt-3.5-turbo

references to being off-duty or unauthorized actions by police officers

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

ulhu

-1.06

 Compass

-0.71

DOI

-0.68

 EntityItem

-0.66

yssey

-0.66

ļéĨĴ

-0.66

ongyang

-0.66

Phi

-0.65

anos

-0.65

illac

-0.64

POSITIVE LOGITS

sized

1.09

level

1.07

bred

1.06

grade

1.05

handed

1.03

touch

1.03

earth

1.02

front

1.02

class

1.01

stage

1.00

Activations Density 0.073%

phrases related to being out of a certain area or off-duty

references to being off-duty or unauthorized actions by police officers

No Comments

No Known Activations