INDEX

Explanations

instructions or descriptions related to physical objects being positioned or located underneath something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating items or situations located beneath something

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

ILY

-0.68

 Begins

-0.65

iterranean

-0.65

arium

-0.63

ä¹ĭ

-0.63

************

-0.62

IER

-0.61

ERO

-0.60

################################

-0.60

HK

-0.59

POSITIVE LOGITS

neath

1.18

whelming

1.08

lining

1.01

pins

1.01

lined

0.96

dogs

0.96

cuts

0.95

valued

0.93

graduate

0.90

cutting

0.90

Activations Density 0.042%

instructions or descriptions related to physical objects being positioned or located underneath something

phrases indicating items or situations located beneath something

No Comments

No Known Activations