INDEX

Explanations

references to the word "Kin" with varying activations indicating different levels of relevance

oai_token-act-pair · gpt-3.5-turbo

references to kin or family relationships

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

shire

-0.79

enance

-0.78

ERE

-0.68

ortium

-0.68

ħĭ

-0.66

ORD

-0.65

Ð¼

-0.65

é¾įå¥ĳå£«

-0.64

orde

-0.64

 incidentally

-0.63

POSITIVE LOGITS

Kin

1.35

kin

1.14

folk

1.13

Kin

1.13

etic

1.11

etics

1.11

etically

1.11

sey

1.04

zie

1.01

ners

0.94

Activations Density 0.011%

references to the word "Kin" with varying activations indicating different levels of relevance

references to kin or family relationships

No Comments

No Known Activations