INDEX

Explanations

phrases related to opting out or unsubscribing from services or updates

oai_token-act-pair · gpt-3.5-turbo

references to opting out of services or subscriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Invention

-0.65

inav

-0.61

 Lack

-0.57

ologne

-0.56

 Ethiop

-0.56

bugs

-0.55

ilts

-0.55

FIX

-0.55

uala

-0.54

enegger

-0.53

POSITIVE LOGITS

ersed

0.65

 altogether

0.64

 anytime

0.63

at

0.59

xon

0.59

scribe

0.58

renheit

0.58

 subscribing

0.58

lect

0.57

 Privacy

0.57

Activations Density 0.016%

phrases related to opting out or unsubscribing from services or updates

references to opting out of services or subscriptions

No Comments

No Known Activations