INDEX

Explanations

references to the name "Rom" followed by different suffixes and word forms

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Rom" or variations thereof, likely indicating a focus on the city of Rome or related subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

gettable

-0.74

umption

-0.73

pressed

-0.71

insula

-0.67

ramid

-0.67

xual

-0.67

 graded

-0.66

raviolet

-0.66

ntil

-0.64

 attendant

-0.64

POSITIVE LOGITS

Rom

1.23

ijn

0.99

antics

0.88

ulus

0.84

Rom

0.84

athan

0.83

agna

0.80

anes

0.80

antically

0.78

aji

0.76

Activations Density 0.008%

references to the name "Rom" followed by different suffixes and word forms

mentions of the term "Rom" or variations thereof, likely indicating a focus on the city of Rome or related subjects

No Comments

No Known Activations