INDEX

Explanations

phrases or terms that are officially recognized or assigned a specific status

oai_token-act-pair · gpt-3.5-turbo

the term "designated" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

RESULTS

-0.80

ebook

-0.79

conom

-0.73

-----------

-0.68

VIDEO

-0.68

QU

-0.67

ears

-0.66

CV

-0.66

Fish

-0.66

co

-0.65

POSITIVE LOGITS

 designated

1.27

 satell

0.98

 designate

0.97

 designation

0.94

 warr

0.84

icter

0.81

 naming

0.80

 citiz

0.79

paces

0.78

anamo

0.76

Activations Density 0.009%

phrases or terms that are officially recognized or assigned a specific status

the term "designated" in various contexts

No Comments

No Known Activations