INDEX

Explanations

news articles or reports about deadly incidents or casualties

oai_token-act-pair · gpt-3.5-turbo

references to individuals who have been killed

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

wcsstore

-1.04

hist

-0.87

 stru

-0.81

soDeliveryDate

-0.75

channelAvailability

-0.74

soType

-0.74

æ©Ł

-0.73

DragonMagazine

-0.72

Han

-0.71

yrinth

-0.68

POSITIVE LOGITS

 spree

0.94

 gunshot

0.84

 rampage

0.79

bolt

0.78

 Trayvon

0.78

 Sandra

0.75

 psychiat

0.75

hyde

0.74

 indict

0.70

 stabbing

0.70

Activations Density 0.100%

news articles or reports about deadly incidents or casualties

references to individuals who have been killed

No Comments

No Known Activations