INDEX

Explanations

expressions of inappropriateness or dissatisfaction regarding social behavior and comments

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

inappropriate, improper, or insensitive

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

 apaixon

-0.34

 económicas

-0.33

 económica

-0.33

 chargeur

-0.33

 economía

-0.33

 remplissage

-0.32

 naviguant

-0.32

 permukaan

-0.32

 cerâmica

-0.32

KommentareTeilen

-0.31

POSITIVE LOGITS

ſp

0.56

 Reſ

0.52

ſelf

0.52

 Majefty

0.52

keted

0.51

 leaſt

0.51

 Chriftian

0.50

 ſei

0.50

 ſta

0.50

 miſ

0.50

Activations Density 0.027%

expressions of inappropriateness or dissatisfaction regarding social behavior and comments

inappropriate, improper, or insensitive

No Comments

No Known Activations