INDEX

Explanations

expressions of strong negative emotions, particularly anger and frustration

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

expressions of anger or rage

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

bootstrapcdn

-0.58

 tiguan

-0.56

 ErrIntOverflow

-0.56

HasAnnotation

-0.55

 ligiloj

-0.55

Datuak

-0.54

:✨

-0.53

EndGlobalSection

-0.52

 nahilalakip

-0.51

 tartalomajánló

-0.51

POSITIVE LOGITS

 angry

0.79

 colère

0.73

 anger

0.69

 annoyed

0.69

angry

0.68

Angry

0.67

 angered

0.67

 indignant

0.66

怒

0.66

 furious

0.65

Activations Density 0.027%

expressions of strong negative emotions, particularly anger and frustration

expressions of anger or rage

No Comments

No Known Activations

expressions of strong negative emotions, particularly anger and frustration

expressions of anger or rage

No Comments

No Known Activations