INDEX

Explanations

instances of observing or visualizing actions and events

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

watching people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

yntaxException

-0.67

 GenerationType

-0.64

StoryboardSegue

-0.63

fjspx

-0.63

majánló

-0.63

 HttpNotFound

-0.62

medriver

-0.61

 actionMode

-0.60

AsUp

-0.60

 indígen

-0.59

POSITIVE LOGITS

 watching

0.47

watching

0.40

 Watching

0.40

 beautiful

0.34

 passers

0.34

 Watch

0.34

watch

0.33

 watch

0.32

pass

0.32

 WATCH

0.31

Activations Density 0.011%

instances of observing or visualizing actions and events

watching people

No Comments

No Known Activations