INDEX

Explanations

references to different issues and topics in society or the world

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

CLASSIFIED

-0.79

oola

-0.71

ALSE

-0.70

NOW

-0.69

HAEL

-0.67

âĿ

-0.67

 caut

-0.66

USB

-0.65

)</

-0.63

=""

-0.63

POSITIVE LOGITS

 lieu

1.23

efficiency

1.21

 accordance

1.21

effic

1.21

clusions

1.16

 conjunction

1.16

 favor

1.10

 vitro

1.09

versions

1.05

ordinate

1.03

Activations Density 0.686%

references to different issues and topics in society or the world

No Comments

No Known Activations