INDEX

Explanations

proper nouns from the entertainment industry and possibly legal or political domains

oai_token-act-pair · gpt-3.5-turbo

references to popular media, specifically related to titles and characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

".[

-0.65

 unfavorable

-0.55

Compat

-0.53

Ĥİ

-0.52

'."

-0.51

 favour

-0.51

?".

-0.50

 favourable

-0.50

 deterrent

-0.47

',"

-0.47

POSITIVE LOGITS

aughters

0.58

anky

0.55

çĶ°

0.54

reprene

0.51

olphin

0.50

ilingual

0.50

 pseudonym

0.49

lore

0.49

esan

0.49

itled

0.49

Activations Density 1.773%

proper nouns from the entertainment industry and possibly legal or political domains

references to popular media, specifically related to titles and characters

No Comments

No Known Activations