INDEX

Explanations

names followed by 'Jr' as a suffix

oai_token-act-pair · gpt-3.5-turbo

the suffix "Jr" indicating junior titles associated with individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

=-=-=-=-=-=-=-=-

-0.78

Front

-0.71

 è£ıç

-0.70

keeping

-0.68

graduate

-0.67

Cover

-0.66

 Marketable

-0.64

BACK

-0.63

point

-0.63

 Petraeus

-0.62

POSITIVE LOGITS

onse

1.02

Jr

0.99

.,

0.90

igne

0.88

Sr

0.86

inx

0.86

uez

0.85

arcer

0.85

usalem

0.78

quez

0.76

Activations Density 0.015%

names followed by 'Jr' as a suffix

the suffix "Jr" indicating junior titles associated with individuals

No Comments

No Known Activations