INDEX

Explanations

expressions indicating win-win scenarios or mutual benefits

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

"win" or "wins"

np_acts-logits-general · gemini-2.0-flash

win or wins

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 ModelExpression

-0.87

WithMany

-0.64

itinéraire

-0.60

Sucesor

-0.59

úgó

-0.59

OGND

-0.58

Diwedd

-0.57



-0.54

__':

-0.53

ůli

-0.53

POSITIVE LOGITS

win

2.33

win

1.96

Win

1.86

Win

1.82

WIN

1.75

WIN

1.73

 wins

1.53

 winning

1.33

 winner

1.29

wins

1.27

Activations Density 0.119%

expressions indicating win-win scenarios or mutual benefits

"win" or "wins"

win or wins

No Comments

No Known Activations