INDEX

Explanations

variations of the word "accusation."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

geist

-0.68

SG

-0.66

wich

-0.64

NING

-0.64

bian

-0.63

 Sawyer

-0.63

OPLE

-0.62

çķ

-0.61

lyak

-0.61

lda

-0.60

POSITIVE LOGITS

ompl

1.41

redited

1.39

urate

1.35

uracy

1.31

identally

1.30

enture

1.25

idental

1.25

laim

1.21

used

1.20

ustomed

1.19

Activations Density 0.011%

variations of the word "accusation."

No Comments

No Known Activations