INDEX

Explanations

References to individuals being arrested

oai_token-act-pair · gpt-3.5-turbo

instances of the word "arrested" in various contexts

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

 enthus

-0.77

vironment

-0.75

emonic

-0.73

morrow

-0.70

clerosis

-0.68

 Remastered

-0.68

lag

-0.67

WE

-0.65

urgical

-0.65

 preference

-0.64

POSITIVE LOGITS

 arrested

0.99

 arrest

0.93

 arrests

0.85

 apprehended

0.83

 wiret

0.76

nab

0.75

 handcuffed

0.74

 detained

0.72

 Arrest

0.71

 decriminal

0.69

Activations Density 0.018%

References to individuals being arrested

instances of the word "arrested" in various contexts

No Comments

No Known Activations