INDEX

Explanations

the word "Em" followed by a number, particularly "Em9" or "Em10."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Em."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Kubrick

-0.82

 Icelandic

-0.73

 Idle

-0.71

 Elves

-0.71

ãģį

-0.70

 Iceland

-0.70

aird

-0.69

WAYS

-0.68

ASC

-0.67

 sund

-0.66

POSITIVE LOGITS

phasis

1.31

otions

1.28

otional

1.28

igrant

1.23

otion

1.19

igrants

1.15

brace

1.14

issions

1.14

ission

1.14

blems

1.12

Activations Density 0.013%

the word "Em" followed by a number, particularly "Em9" or "Em10."

occurrences of the term "Em."

No Comments

No Known Activations