INDEX

Explanations

phrases related to adding or introducing something new or additional

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the addition or introduction of new information or perspectives

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

atcher

-0.65

-0.64

Application

-0.62

CBC

-0.60

OSED

-0.60

published

-0.60

Win

-0.59

EStream

-0.59

issued

-0.58

 PAGE

-0.58

POSITIVE LOGITS

 flair

0.98

 layer

0.89

endum

0.87

 thereto

0.84

 dots

0.82

 layers

0.81

 onto

0.80

 reinforcements

0.77

 insult

0.75

 extra

0.74

Activations Density 0.250%

phrases related to adding or introducing something new or additional

phrases that indicate the addition or introduction of new information or perspectives

No Comments

No Known Activations