INDEX

Explanations

phrases that indicate uncertainty or tentative statements

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

Na and На prefixes

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

<h3>

-0.55

Esprit

-0.50

-0.48

 Erfurt

-0.47

 Jambi

-0.47

 Esprit

-0.44

íes

-0.44

<h1>

-0.42

 Frucht

-0.42

 skjø

-0.41

POSITIVE LOGITS

Na

0.73

Na

0.72

na

0.70

 setName

0.69

NA

0.66

NDA

0.66

На

0.64

 onStart

0.64

na

0.63

 lastname

0.63

Activations Density 0.009%

phrases that indicate uncertainty or tentative statements

Na and На prefixes

No Comments

No Known Activations