INDEX

Explanations

words related to technical instructions and procedures

oai_token-act-pair · gpt-3.5-turbo

terms related to digital content management and organization

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 intolerable

-0.71

 unfit

-0.68

 Vietnam

-0.65

heid

-0.65

 Parenthood

-0.65

 Colossus

-0.64

 Fathers

-0.63

 Patriot

-0.63

 Accord

-0.63

 Incarn

-0.63

POSITIVE LOGITS

links

1.02

HTML

1.01

 bookmark

0.96

code

0.88

pages

0.87

 formatting

0.86

codes

0.86

 notation

0.84

isode

0.84

FAQ

0.84

Activations Density 0.721%

words related to technical instructions and procedures

terms related to digital content management and organization

No Comments

No Known Activations