INDEX

Explanations

words related to magnetic forces and specific names like "Magnet" and "Xavier"

oai_token-act-pair · gpt-3.5-turbo

references to the character Magneto and the concept of magnetism

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãģį

-0.92

OWS

-0.78

 Citiz

-0.67

à©

-0.64

ORD

-0.63

UD

-0.61

 Applicant

-0.58

 center

-0.56

razen

-0.56

thought

-0.56

POSITIVE LOGITS

ophon

1.03

stadt

0.99

eers

0.98

xon

0.97

ical

0.95

istical

0.94

bol

0.92

eer

0.90

hester

0.89

ertodd

0.87

Activations Density 0.052%

words related to magnetic forces and specific names like "Magnet" and "Xavier"

references to the character Magneto and the concept of magnetism

No Comments

No Known Activations