INDEX

Explanations

conjunctions and linking words that suggest relationships or dependencies in complex sentences

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

Personendaten

-0.68

transQ

-0.63

BibitemShut

-0.62

 geweſen

-0.61

 müſſen

-0.61

<pad>

-0.61

bildtitel

-0.61

 ſeines

-0.61

 ſeyn

-0.61

<unused43>

-0.61

POSITIVE LOGITS

 thus

0.40

 hence

0.37

 således

0.33

 cioè

0.33

 nagu

0.32

 daardoor

0.31

それで

0.31

しかも

0.31

 perciò

0.31

 Thus

0.30

Activations Density 0.454%

conjunctions and linking words that suggest relationships or dependencies in complex sentences

No Comments

No Known Activations