INDEX

Explanations

information related to past and current occupations or positions held by individuals

oai_token-act-pair · gpt-3.5-turbo

references to individuals' current and past employment or roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 rejecting

-0.72

vine

-0.71

immune

-0.70

bidden

-0.70

yip

-0.69

uri

-0.68

 Saying

-0.66

peat

-0.66

happy

-0.65

apest

-0.65

POSITIVE LOGITS

 oversaw

1.36

 oversees

1.33

 oversee

1.24

 superv

1.21

 worked

1.18

 coached

1.15

 serves

1.15

 owns

1.14

 coordinates

1.13

 supervised

1.13

Activations Density 0.321%

information related to past and current occupations or positions held by individuals

references to individuals' current and past employment or roles

No Comments

No Known Activations