INDEX

Explanations

instances of manipulation and emotional conflict in relationships

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

recklessly inflicting harm or teasing

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

words and tokens that express strong emotion, judgment, or emphatic emphasis (opinionated/emotional language).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

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

 ujednoznacz

-0.46

Atsauces

-0.42

 MainAxisSize

-0.40

 esternos

-0.39

észetes

-0.39

gridx

-0.38

 stránka

-0.38

脚注の使い方

-0.38

ataire

-0.38

 fija

-0.37

POSITIVE LOGITS

évaluateur

0.46

 tease

0.44

Brin

0.44

責任

0.43

 mischie

0.43

 inflicting

0.42

 teasing

0.42

 irresponsible

0.42

alsch

0.41

omin

0.41

Activations Density 0.084%

instances of manipulation and emotional conflict in relationships

recklessly inflicting harm or teasing

words and tokens that express strong emotion, judgment, or emphatic emphasis (opinionated/emotional language).

No Comments

No Known Activations