INDEX

Explanations

phrases related to individuals in prominent positions or roles

oai_token-act-pair · gpt-3.5-turbo

references to political candidates and their standings

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

 corpus

-0.60

»Ĵ

-0.60

 Kard

-0.60

ibly

-0.59

 Twain

-0.58

 mediation

-0.58

 Temporary

-0.58

insk

-0.58

terday

-0.57

 Brune

-0.57

POSITIVE LOGITS

runner

1.20

facing

1.13

office

1.10

runners

1.03

running

1.02

eyed

0.99

line

0.98

lining

0.96

eye

0.92

Runner

0.92

Activations Density 0.033%

phrases related to individuals in prominent positions or roles

references to political candidates and their standings

No Comments

No Known Activations