INDEX

Explanations

conditional statements and questions directed at the reader

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 33-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.33.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

probably

-0.46

 probably

-0.46

Probably

-0.45

 Probably

-0.43

endregion

-0.40

可能是

-0.39

 Either

-0.37

似乎

-0.36

 نه

-0.36

 zwar

-0.36

POSITIVE LOGITS

 hadn

0.63

 compare

0.61

 truly

0.61

 ever

0.59

 properly

0.56

 Properly

0.56

 jemals

0.54

 malheur

0.54

truly

0.53

 verdaderamente

0.53

Activations Density 0.281%

conditional statements and questions directed at the reader

No Comments

No Known Activations

conditional statements and questions directed at the reader

No Comments

No Known Activations