INDEX

Explanations

phrases indicating attempts to try or sample different options

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/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.34.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

 kasarigan

-0.53

AsUp

-0.49

 preocupación

-0.38

intended

-0.38

⏜

-0.37

フィシャル

-0.37

这座

-0.37

richTextPanel

-0.37

bindingNavigator

-0.36

 preocupação

-0.36

POSITIVE LOGITS

 trial

2.63

 Trial

2.25

trial

2.25

Trial

2.16

 TRIAL

2.05

 trials

2.05

try

1.88

试

1.87

 Trials

1.80

試

1.80

Activations Density 0.676%

phrases indicating attempts to try or sample different options

No Comments

No Known Activations