INDEX

Explanations

mentions of the abbreviation "sn" followed by a number, potentially referencing sarcasm or negative comments in the text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "sn"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

heid

-0.88

xual

-0.77

 PowerPoint

-0.75

mine

-0.72

heit

-0.69

minus

-0.64

EMENT

-0.64

shire

-0.63

minent

-0.62

âĹ¼

-0.62

POSITIVE LOGITS

agged

1.04

obb

1.04

agging

1.04

atching

1.00

ugg

1.00

opes

0.97

atches

0.97

ips

0.95

uggle

0.95

apper

0.94

Activations Density 0.007%

mentions of the abbreviation "sn" followed by a number, potentially referencing sarcasm or negative comments in the text

occurrences of the prefix "sn"

No Comments

No Known Activations