INDEX

Explanations

expressions of remorse or apology

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

Following verbs like "should," "they," or brands like "Volkswagen"

np_acts-logits-general · gemini-2.0-flash

admitting mistakes or apologizing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 30-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_30/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.30.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

醐

-0.40

 dreaming

-0.33

 acquis

-0.32

 soñar

-0.32

 déco

-0.32

 vencer

-0.31

elni

-0.30

 Durand

-0.30

hydro

-0.29

iterrane

-0.29

POSITIVE LOGITS

 apology

1.84

 apologized

1.75

 apologised

1.65

 apologizing

1.65

 apologize

1.64

 apologise

1.55

 apologies

1.55

 apolog

1.41

道歉

1.29

 apologe

1.23

Activations Density 0.483%

expressions of remorse or apology

Following verbs like "should," "they," or brands like "Volkswagen"

admitting mistakes or apologizing

No Comments

No Known Activations