INDEX

Explanations

mentions of authority or power dynamics

oai_token-act-pair · gpt-3.5-turbo

instances of the word "has" and its variations in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Pair

-0.68

 filming

-0.63

dot

-0.63

 recalls

-0.62

etter

-0.62

TG

-0.61

etting

-0.59

burse

-0.59

hooting

-0.58

umping

-0.57

POSITIVE LOGITS

 been

1.27

 gotten

1.05

been

1.03

 behaved

1.02

 gone

1.01

 become

0.99

 begun

0.96

oken

0.94

 ceased

0.93

 fallen

0.93

Activations Density 0.378%

mentions of authority or power dynamics

instances of the word "has" and its variations in context

No Comments

No Known Activations