INDEX

Explanations

mentions of physical injuries or blood-related terms

oai_token-act-pair · gpt-3.5-turbo

instances of the word "bleeding" and related terms

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

å§«

-0.81

wcsstore

-0.75

bnb

-0.74

adal

-0.74

idable

-0.73

BOOK

-0.69

essee

-0.69

inqu

-0.68

lator

-0.68

fman

-0.66

POSITIVE LOGITS

 wounds

0.93

 hemorrh

0.89

 throats

0.83

 blood

0.80

 prof

0.79

 necks

0.78

 bleeding

0.78

 wrists

0.77

 crimson

0.76

 bleed

0.76

Activations Density 0.077%

mentions of physical injuries or blood-related terms

instances of the word "bleeding" and related terms

No Comments

No Known Activations