INDEX

Explanations

action-related words like "want," "achieve," and "perform."

oai_token-act-pair · gpt-3.5-turbo

expressions related to decision-making and personal preferences

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

ushes

-0.79

inently

-0.65

arkable

-0.64

rir

-0.64

 fals

-0.62

 innocence

-0.62

ndra

-0.61

 Freak

-0.61

culosis

-0.61

mone

-0.60

POSITIVE LOGITS

 next

1.13

 depending

1.00

 hereafter

0.95

 depends

0.91

?,

0.91

 tomorrow

0.89

 Next

0.86

 NEXT

0.84

 versus

0.81

FontSize

0.79

Activations Density 0.471%

action-related words like "want," "achieve," and "perform."

expressions related to decision-making and personal preferences

No Comments

No Known Activations