INDEX

Explanations

phrases related to police actions or use of force

oai_token-act-pair · gpt-3.5-turbo

phrases associated with law enforcement actions and incidents

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

isSpecialOrderable

-0.86

hai

-0.82

sonian

-0.79

ãĤ·ãĥ£

-0.77

 UNESCO

-0.76

RET

-0.76

 Maiden

-0.76

aceae

-0.75

 realise

-0.74

dream

-0.73

POSITIVE LOGITS

 unarmed

0.96

 misdemeanor

0.95

 homicide

0.93

 lawful

0.92

 arrest

0.92

 felon

0.92

 felony

0.91

 policing

0.90

 firearm

0.90

 offender

0.90

Activations Density 0.319%

phrases related to police actions or use of force

phrases associated with law enforcement actions and incidents

No Comments

No Known Activations