INDEX

Explanations

the word "any"

oai_token-act-pair · gpt-3.5-turbo

phrases that generalize situations or events

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

rex

-0.76

endas

-0.74

combe

-0.69

icz

-0.68

rez

-0.68

ÃŁ

-0.67

appings

-0.64

seless

-0.63

staking

-0.62

lights

-0.62

POSITIVE LOGITS

THING

1.12

 conceivable

1.05

body

0.88

WHERE

0.87

 imaginable

0.87

 particular

0.83

 sort

0.78

ONE

0.77

 given

0.77

 foreseeable

0.77

Activations Density 0.069%

the word "any"

phrases that generalize situations or events

No Comments

No Known Activations