INDEX

Explanations

words related to medical conditions or processes

oai_token-act-pair · gpt-3.5-turbo

instances of the prefix "en-" used in words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

DragonMagazine

-0.73

TPPStreamerBot

-0.72

ageddon

-0.71

 Heights

-0.70

 Dare

-0.69

ĳå£«

-0.67

 Orbit

-0.67

tone

-0.67

Berry

-0.66

 Tanks

-0.65

POSITIVE LOGITS

viron

1.27

 masse

1.04

en

1.01

cyclopedia

0.99

semb

0.97

chanted

0.93

ounter

0.92

joined

0.91

forcer

0.91

emy

0.88

Activations Density 0.008%

words related to medical conditions or processes

instances of the prefix "en-" used in words

No Comments

No Known Activations