INDEX

Explanations

news headlines that include additional information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "More" indicating ongoing or additional information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

uca

-0.78

atan

-0.71

keeping

-0.70

ovie

-0.69

xtap

-0.69

Fram

-0.67

idated

-0.67

DragonMagazine

-0.66

ãĥ¼ãĥĨ

-0.66

OCK

-0.64

POSITIVE LOGITS

 importantly

1.24

 than

1.08

ado

1.07

 Than

1.07

 important

0.79

 Likely

0.78

 interesting

0.76

 sophisticated

0.76

 troubling

0.73

 worrisome

0.73

Activations Density 0.067%

news headlines that include additional information

instances of the word "More" indicating ongoing or additional information

No Comments

No Known Activations