INDEX

Explanations

references to lawyers or legal proceedings

oai_token-act-pair · gpt-3.5-turbo

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

 Reincarn

-0.82

plings

-0.71

obyl

-0.68

Poll

-0.66

 Memories

-0.65

 Marines

-0.65

 Railway

-0.65

efully

-0.64

 Territories

-0.63

Neh

-0.63

POSITIVE LOGITS

 Gloria

0.99

 representing

0.92

 specializing

0.88

 solicitor

0.86

yer

0.83

 suing

0.82

 attorney

0.80

suit

0.79

yers

0.79

maid

0.78

Activations Density 0.052%

references to lawyers or legal proceedings

No Comments

No Known Activations