INDEX

Explanations

phrases related to unhealthy food items

oai_token-act-pair · gpt-3.5-turbo

references to junk food or items labeled as "junk."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

-0.85

cing

-0.72

ndum

-0.66

 Lauder

-0.65

BILITY

-0.65

 Wilkinson

-0.65

resp

-0.64

TEXTURE

-0.63

gyn

-0.63

 Scarlett

-0.62

POSITIVE LOGITS

 junk

1.12

 heap

0.93

 dumps

0.87

 drawer

0.81

 pile

0.79

raft

0.78

rats

0.78

abies

0.78

unks

0.77

 mail

0.77

Activations Density 0.005%

phrases related to unhealthy food items

references to junk food or items labeled as "junk."

No Comments

No Known Activations