INDEX

Explanations

names of people

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names related to people

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

 youtube

-0.63

 Birthday

-0.63

venge

-0.61

Offline

-0.61

Bound

-0.61

renheit

-0.60

 Fortress

-0.57

 starving

-0.57

 Thumbnails

-0.55

IVE

-0.54

POSITIVE LOGITS

 added

1.43

 said

1.39

 explained

1.32

 cautioned

1.24

 noted

1.22

 argued

1.20

 pointed

1.17

 speculated

1.14

 recalled

1.14

 conceded

1.12

Activations Density 0.104%

names of people

proper nouns and names related to people

No Comments

No Known Activations