INDEX

Explanations

words related to horror, violence, and crime

oai_token-act-pair · gpt-3.5-turbo

references to violent or traumatic events

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

pai

-0.82

arten

-0.78

sers

-0.77

ership

-0.73

rase

-0.71

essor

-0.69

sama

-0.69

sterdam

-0.68

cius

-0.67

gat

-0.67

POSITIVE LOGITS

 atrocities

0.90

 horrors

0.89

 ordeal

0.81

 earthqu

0.79

 asylum

0.78

 nightmares

0.77

ally

0.77

 Revelations

0.76

 imagery

0.76

 tales

0.75

Activations Density 0.045%

words related to horror, violence, and crime

references to violent or traumatic events

No Comments

No Known Activations