INDEX

Explanations

scenarios or situations posed as questions, often beginning with "What if" or "What if we"

oai_token-act-pair · gpt-3.5-turbo

conditional questions or scenarios presented with "what if."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

igmatic

-0.79

cedented

-0.77

vant

-0.77

cised

-0.73

abre

-0.73

nect

-0.72

 pione

-0.71

enfranch

-0.69

ply

-0.69

POSITIVE LOGITS

 someday

1.06

 somebody

0.90

 someone

0.88

...?

0.86

we

0.80

 they

0.79

 there

0.78

you

0.78

 somehow

0.77

 instead

0.76

Activations Density 0.063%

scenarios or situations posed as questions, often beginning with "What if" or "What if we"

conditional questions or scenarios presented with "what if."

No Comments

No Known Activations