INDEX

Explanations

specific mentions of the word "ior"

oai_token-act-pair · gpt-3.5-turbo

references to hierarchical relationships or titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

lished

-1.02

hov

-0.76

ting

-0.72

isSpecialOrderable

-0.72

alogue

-0.68

soDeliveryDate

-0.67

apo

-0.64

puter

-0.63

ãĥīãĥ©

-0.63

RGB

-0.63

POSITIVE LOGITS

ior

1.30

gio

1.10

IOR

1.01

ity

0.84

iculture

0.80

idad

0.80

ITY

0.79

acies

0.78

iors

0.77

andom

0.75

Activations Density 0.008%

specific mentions of the word "ior"

references to hierarchical relationships or titles

No Comments

No Known Activations