INDEX

Explanations

ellipses or incomplete thoughts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

市镇

-0.51

ally

-0.47

 itself

-0.44

 utafitiHapana

-0.44

matically

-0.44

 raiſ

-0.43

 grand

-0.42

PREFERRED

-0.39

ATORY

-0.39

…

-0.38

POSITIVE LOGITS

.;

0.49

RenderAtEndOf

0.47

httphttps

0.47

 initComponents

0.46

 ...'

0.46

endwhile

0.46

FieldNumber

0.45

.:

0.45

...

0.45

 laikā

0.45

Activations Density 0.277%

ellipses or incomplete thoughts

No Comments

No Known Activations