INDEX

Explanations

references to sheriff's offices

oai_token-act-pair · gpt-3.5-turbo

references to law enforcement, specifically the term "Sheriff."

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

lihood

-0.83

ctory

-0.73

umar

-0.71

aeda

-0.69

auri

-0.67

tar

-0.66

ces

-0.65

 raining

-0.65

arial

-0.64

jad

-0.63

POSITIVE LOGITS

 Sheriff

1.44

 sheriff

1.14

 Marshal

1.03

Sher

0.89

 Supervisor

0.88

 deputies

0.85

 Arpaio

0.84

iffs

0.80

osate

0.80

 sher

0.79

Activations Density 0.012%

references to sheriff's offices

references to law enforcement, specifically the term "Sheriff."

No Comments

No Known Activations