INDEX

Explanations

proper nouns of names of people and places

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 himself

-0.80

 erection

-0.75

GoldMagikarp

-0.67

ATIONS

-0.64

 looph

-0.64

 tyre

-0.63

 hindsight

-0.63

 rall

-0.63

uesday

-0.62

Sax

-0.62

POSITIVE LOGITS

 herself

1.22

ova

1.11

kaya

0.91

 bikini

0.88

eva

0.86

 miscar

0.84

 husband

0.84

aura

0.84

 maid

0.84

 veil

0.83

Activations Density 0.423%

proper nouns of names of people and places

No Comments

No Known Activations