INDEX

Explanations

references to specific individuals or groups in contexts of trust and opinion

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

tolerated condition

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

Hentet

-0.49

itake

-0.46

ftagPool

-0.45

awtextra

-0.44

parsedMessage

-0.43

contentLoaded

-0.43

 مشارکت‌کنندگان

-0.42

gardens

-0.42

holi

-0.42

 Qaraldi

-0.41

POSITIVE LOGITS

 المعيارى

0.51

 perverse

0.46

 Toler

0.43

Toler

0.40

లాలు

0.38

len

0.37

 tolerated

0.35

nestjs

0.35

 compromised

0.35

ophiles

0.35

Activations Density 0.220%

references to specific individuals or groups in contexts of trust and opinion

tolerated condition

No Comments

No Known Activations