INDEX

Explanations

terms related to verbal attacks and criticisms

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

attacks against political opponents

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

thâu

-0.73

 informée

-0.70

ंदीखरीदारी

-0.59

GEBURTSDATUM

-0.57

省市镇

-0.49

 surla

-0.48

HasAnnotation

-0.47

 Мексичка

-0.47

Vidite

-0.46

Демографія

-0.46

POSITIVE LOGITS

 attack

0.57

 attacking

0.56

 Attack

0.56

attack

0.54

 attacks

0.52

 criticism

0.51

Attack

0.47

 criticizing

0.47

ATTACK

0.47

 ataques

0.46

Activations Density 0.044%

terms related to verbal attacks and criticisms

attacks against political opponents

No Comments

No Known Activations