INDEX

Explanations

expressions related to data structure elements or values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.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

Tikang

-0.89

<unused68>

-0.86

<unused52>

-0.86

𑄢

-0.85

<unused41>

-0.85

<unused23>

-0.85

<unused79>

-0.85

𑄣

-0.85

[@BOS@]

-0.85

<unused8>

-0.85

POSITIVE LOGITS

<eos>

0.56

 Again

0.53

...

0.50

 again

0.49

 Finally

0.46

Again

0.46

↵↵

0.42

Finally

0.38

...

0.37

 Further

0.37

Activations Density 0.904%

expressions related to data structure elements or values

No Comments

No Known Activations