INDEX

Explanations

phrases related to actions of violence, particularly killing

oai_token-act-pair · gpt-3.5-turbo

the word "kill" in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arity

-0.73

anwhile

-0.72

 concess

-0.72

 Collider

-0.69

umn

-0.67

rial

-0.66

Jer

-0.66

ourn

-0.65

 Hilbert

-0.65

MpServer

-0.65

POSITIVE LOGITS

 kill

0.97

 spree

0.89

 kills

0.84

hog

0.84

 killers

0.84

 killing

0.83

icides

0.83

killer

0.83

icidal

0.81

houses

0.78

Activations Density 0.020%

phrases related to actions of violence, particularly killing

the word "kill" in various contexts

No Comments

No Known Activations