INDEX

Explanations

references to physical doors

oai_token-act-pair · gpt-3.5-turbo

references to various types of doors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ollah

-0.71

TING

-0.70

 Hots

-0.68

udeb

-0.68

uke

-0.67

lihood

-0.67

ousand

-0.66

ting

-0.65

 PubMed

-0.64

ovych

-0.64

POSITIVE LOGITS

bell

1.27

door

1.21

steps

1.20

doors

1.05

 doors

1.00

plates

0.98

posts

0.96

ways

0.92

holes

0.92

 opener

0.92

Activations Density 0.029%

references to physical doors

references to various types of doors

No Comments

No Known Activations