INDEX

Explanations

information about blogs or posts

oai_token-act-pair · gpt-3.5-turbo

sequences of formatting or structural elements, such as dividers or placeholders

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

pex

-0.68

Peb

-0.64

bery

-0.64

 Azerb

-0.63

forth

-0.62

shed

-0.62

chal

-0.61

selves

-0.61

Gaw

-0.60

 Nost

-0.59

POSITIVE LOGITS

 âĢº

0.81

][

0.79

0.78

 thumbnail

0.73

 âĸº

0.72

 partName

0.69

quickShipAvailable

0.69

enter

0.69

 guiActiveUnfocused

0.69

 Container

0.67

Activations Density 0.238%

information about blogs or posts

sequences of formatting or structural elements, such as dividers or placeholders

No Comments

No Known Activations