INDEX

Explanations

phrases with a specific pattern of letters and numbers, possibly related to names or codes

oai_token-act-pair · gpt-3.5-turbo

specific proper nouns, likely related to names, brands, or organizations

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

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

×©

-0.73

Liter

-0.69

Favorite

-0.67

RIP

-0.66

ãĥİ

-0.66

 liking

-0.63

Ü

-0.63

URRENT

-0.60

 Thumbnails

-0.60

soDeliveryDate

-0.60

POSITIVE LOGITS

 said

1.19

 says

1.18

 concedes

1.15

 acknowledges

1.14

 argues

1.12

 explains

1.11

 explained

1.11

 contends

1.06

 concludes

1.05

 admits

1.05

Activations Density 0.323%

phrases with a specific pattern of letters and numbers, possibly related to names or codes

specific proper nouns, likely related to names, brands, or organizations

No Comments

No Known Activations