INDEX

Explanations

information that includes numbers within parentheses, specifically high activation on (10) and (9) along with some (8) activations

oai_token-act-pair · gpt-3.5-turbo

multiple parentheses in a scientific or analytical context

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

 Lumpur

-0.76

 lull

-0.71

quished

-0.69

icy

-0.65

 tempted

-0.65

 glare

-0.64

 swoop

-0.64

 ally

-0.63

 reneg

-0.63

 beaten

-0.62

POSITIVE LOGITS

see

1.54

including

1.40

Figure

1.34

sic

1.31

emphasis

1.28

excluding

1.27

both

1.24

which

1.23

such

1.23

typically

1.23

Activations Density 0.162%

information that includes numbers within parentheses, specifically high activation on (10) and (9) along with some (8) activations

multiple parentheses in a scientific or analytical context

No Comments

No Known Activations