INDEX

Explanations

phrases related to positions of authority or leadership

oai_token-act-pair · gpt-3.5-turbo

titles and roles related to leadership positions

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

 Ferr

-0.76

 WATCHED

-0.65

yr

-0.63

ENDED

-0.62

asca

-0.62

 probing

-0.61

OTOS

-0.61

 Homes

-0.61

ICAN

-0.61

 subjected

-0.60

POSITIVE LOGITS

chief

1.50

Chief

1.18

doms

1.01

pants

0.89

 Chief

0.82

pin

0.79

thood

0.76

ãĤ¦ãĤ¹

0.74

ãĥĥãĥĪ

0.74

fal

0.72

Activations Density 0.010%

phrases related to positions of authority or leadership

titles and roles related to leadership positions

No Comments

No Known Activations