INDEX

Explanations

news-related contexts or crime events indicated by words such as "arrested."

oai_token-act-pair · gpt-3.5-turbo

instances of arrests

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

casters

-0.80

gran

-0.73

gaard

-0.72

emic

-0.71

ankind

-0.71

issance

-0.70

mology

-0.69

pace

-0.68

clips

-0.67

fax

-0.67

POSITIVE LOGITS

 criminally

0.77

 Officers

0.75

PKK

0.75

 Constable

0.74

 arrested

0.72

 felon

0.71

 Trooper

0.71

 Marshal

0.69

 handcuffed

0.68

 Warrant

0.67

Activations Density 0.052%

news-related contexts or crime events indicated by words such as "arrested."

instances of arrests

No Comments

No Known Activations