INDEX

Explanations

terms related to destruction or violent actions, specifically involving mutilation, decapitation, and massacre

oai_token-act-pair · gpt-3.5-turbo

words associated with violence and severe bodily harm

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ellig

-0.68

 latitude

-0.68

 facult

-0.68

BET

-0.61

 governed

-0.61

soDeliveryDate

-0.60

arten

-0.60

ordan

-0.60

 permitted

-0.59

oother

-0.59

POSITIVE LOGITS

 havoc

0.83

 inflicted

0.81

ãĥĩ

0.78

ĺħ

0.75

DAQ

0.70

ciating

0.70

 Slaughter

0.69

à¨

0.69

adoes

0.68

 Blow

0.67

Activations Density 0.088%

terms related to destruction or violent actions, specifically involving mutilation, decapitation, and massacre

words associated with violence and severe bodily harm

No Comments

No Known Activations