INDEX

Explanations

mentions of editorial content or opinion pieces

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Op" in various contexts

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

wagen

-0.71

 mileage

-0.70

 wart

-0.70

é¾įå¥ĳå£«

-0.69

 devils

-0.64

åŃĲ

-0.63

 Gateway

-0.62

 calves

-0.61

llah

-0.60

è¦ļéĨĴ

-0.60

POSITIVE LOGITS

inion

1.33

aque

1.32

osite

1.27

ulent

1.19

aqu

1.15

ulence

1.14

onent

1.12

yright

1.04

codes

1.04

iates

1.02

Activations Density 0.028%

mentions of editorial content or opinion pieces

occurrences of the word "Op" in various contexts

No Comments

No Known Activations