INDEX

Explanations

conditional expressions and phrases indicating requirements or dependencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 40-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_40/width_16k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.40.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

couz

-0.50

TagMode

-0.49

themselves

-0.46

herself

-0.46

himself

-0.46

Them

-0.45

them

-0.43

 ourselves

-0.42

*/].

-0.42

 honom

-0.41

POSITIVE LOGITS

 there

1.05

 applicable

0.85

 needed

0.80

 they

0.78

 necessary

0.78

 appropriate

0.68

 houver

0.68

 unsure

0.66

 unable

0.66

 someone

0.65

Activations Density 0.524%

conditional expressions and phrases indicating requirements or dependencies

No Comments

No Known Activations