INDEX

Explanations

terms related to technology and instructions

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

worth

-0.83

æĪ¦

-0.78

olid

-0.69

ghazi

-0.67

elong

-0.66

emia

-0.66

iquid

-0.66

 Yugoslavia

-0.65

 Qiao

-0.64

utsche

-0.64

POSITIVE LOGITS

CLI

0.96

 Menu

0.91

 menu

0.91

 shortcut

0.89

QR

0.88

 Ctrl

0.88

 Remote

0.87

 Toggle

0.87

 touchscreen

0.86

 Preferences

0.86

Activations Density 0.256%

terms related to technology and instructions

No Comments

No Known Activations