INDEX

Explanations

mentions of people being taken to hospitals or medical centers

oai_token-act-pair · gpt-3.5-turbo

phrases related to being taken to a hospital or medical facility

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 stripe

-0.71

 hypocrisy

-0.68

 rebell

-0.68

 beginnings

-0.67

 illusions

-0.65

 genesis

-0.64

 artif

-0.63

ggles

-0.63

 fract

-0.63

 thickness

-0.62

POSITIVE LOGITS

 medically

1.01

ospital

0.92

 hospital

0.91

 handcuffed

0.86

 hospitalized

0.86

untarily

0.85

 detained

0.82

 stret

0.80

 safely

0.78

uggage

0.78

Activations Density 0.103%

mentions of people being taken to hospitals or medical centers

phrases related to being taken to a hospital or medical facility

No Comments

No Known Activations