INDEX

Explanations

words related to mortality or endings

oai_token-act-pair · gpt-3.5-turbo

references to mortality or morbid themes

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

BOOK

-0.77

orney

-0.67

XY

-0.67

 sockets

-0.64

payer

-0.63

 peeled

-0.63

ECH

-0.62

 Wheeler

-0.62

sets

-0.61

 Tablet

-0.60

POSITIVE LOGITS

gue

1.08

aceutical

0.99

gage

0.90

mor

0.86

idian

0.83

itary

0.83

iday

0.82

phe

0.82

asses

0.81

ose

0.79

Activations Density 0.008%

words related to mortality or endings

references to mortality or morbid themes

No Comments

No Known Activations