INDEX

Explanations

phrases or words related to decisions or actions to proceed with a plan or activity

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a directive or an authorization related to approval or permissions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Nicarag

-0.70

 Arabian

-0.67

 EntityItem

-0.66

ciating

-0.65

 eyebrow

-0.63

 Vaughn

-0.62

 Fiscal

-0.59

 PLoS

-0.58

HCR

-0.58

 Feder

-0.56

POSITIVE LOGITS

ahead

1.14

around

1.14

through

1.05

back

1.05

away

1.05

cart

1.03

home

0.99

along

0.96

out

0.96

cation

0.94

Activations Density 0.056%

phrases or words related to decisions or actions to proceed with a plan or activity

phrases that indicate a directive or an authorization related to approval or permissions

No Comments

No Known Activations