INDEX

Explanations

mentions of fatalities or deaths in various incidents

oai_token-act-pair · gpt-3.5-turbo

instances of the word "killed" in various contexts of violence or death

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

-0.81

rium

-0.75

Cola

-0.73

DragonMagazine

-0.70

yrinth

-0.70

issue

-0.70

issance

-0.69

arity

-0.69

arist

-0.67

ellation

-0.66

POSITIVE LOGITS

 spree

0.88

 slain

0.72

 bystand

0.72

 bystanders

0.72

 rampage

0.71

ousel

0.71

ĺħ

0.71

abouts

0.67

selage

0.67

locked

0.67

Activations Density 0.026%

mentions of fatalities or deaths in various incidents

instances of the word "killed" in various contexts of violence or death

No Comments

No Known Activations