INDEX

Explanations

instances of names/signatures associated with photographs or images

oai_token-act-pair · gpt-3.5-turbo

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

 opium

-0.87

 Huma

-0.86

idium

-0.85

 Cipher

-0.83

Obj

-0.83

dy

-0.81

 Router

-0.79

 router

-0.78

yip

-0.78

IDF

-0.77

POSITIVE LOGITS

 Scott

2.53

Scott

2.39

 Scot

1.26

Scot

1.13

Mart

1.04

 Gregg

0.99

 Stuart

0.97

 Scottish

0.97

 Martinez

0.96

 Scots

0.95

Activations Density 0.143%

instances of names/signatures associated with photographs or images

No Comments

No Known Activations