INDEX

Explanations

instances of text expressing enjoyment, excitement, and positive experiences

oai_token-act-pair · gpt-3.5-turbo

expressions related to fun and enjoyable activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

axter

-0.72

ortium

-0.72

ãĥ¯ãĥ³

-0.72

ãĥĥãĥĪ

-0.70

 jurisdiction

-0.69

imester

-0.68

 unilaterally

-0.66

 Regulatory

-0.65

 looph

-0.65

insured

-0.65

POSITIVE LOGITS

GIF

0.88

Adventure

0.75

 romance

0.75

gif

0.74

 gigg

0.73

awe

0.72

 trivia

0.71

Enjoy

0.70

 selfies

0.69

 creativity

0.69

Activations Density 0.557%

instances of text expressing enjoyment, excitement, and positive experiences

expressions related to fun and enjoyable activities

No Comments

No Known Activations