INDEX

Explanations

the word "wanna" at various activation levels

oai_token-act-pair · gpt-3.5-turbo

expressions of desire or intention to take action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

å£«

-0.86

VERTISEMENT

-0.84

advertisement

-0.83

arian

-0.76

loo

-0.74

lain

-0.72

idem

-0.71

sequ

-0.70

edience

-0.69

ochond

-0.68

POSITIVE LOGITS

 wanna

1.29

ignt

0.80

nab

0.75

pping

0.74

pload

0.71

 aspire

0.70

 listen

0.70

 hear

0.69

reprene

0.69

ya

0.68

Activations Density 0.005%

the word "wanna" at various activation levels

expressions of desire or intention to take action

No Comments

No Known Activations