INDEX

Explanations

phrases that express commands or suggestions, often starting with "let's."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

 ligiloj

-0.59

 kasarigan

-0.59

msgTypes

-0.56

 otomatig

-0.53

évaluateur

-0.50

 LEYENDO

-0.50

ectoria

-0.48

InjectAttribute

-0.47

 noDo

-0.47

tangentMode

-0.46

POSITIVE LOGITS

 rewind

0.56

 recap

0.56

 briefly

0.53

 digress

0.53

 quickly

0.53

 suppose

0.51

 eens

0.51

 dissect

0.49

 pause

0.48

 assume

0.47

Activations Density 0.231%

phrases that express commands or suggestions, often starting with "let's."

No Comments

No Known Activations