INDEX

Explanations

phrases related to membership and sign-up prompts

oai_token-act-pair · gpt-3.5-turbo

calls to action related to signing up for a service

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

©¶æ

-0.75

 QuÃ©

-0.72

 Siem

-0.71

 dressing

-0.67

rador

-0.65

ovember

-0.65

 Hann

-0.65

 mosqu

-0.63

————

-0.61

wives

-0.61

POSITIVE LOGITS

rawdownloadcloneembedreportprint

1.29

 unlocks

1.01

 Subtle

0.94

 Pastebin

0.80

ilded

0.72

 Hover

0.69

Detailed

0.68

 pledged

0.65

hidden

0.64

lvl

0.63

Activations Density 0.020%

phrases related to membership and sign-up prompts

calls to action related to signing up for a service

No Comments

No Known Activations