INDEX

Explanations

phrases related to user rights and content moderation

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

remove or refuse

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

fjspx

-0.67

尽

-0.44

OGND

-0.43

 Pick

-0.40

Dil

-0.40

解

-0.40

 nakalista

-0.39

COMPAR

-0.38

 kend

-0.38

тельству

-0.38

POSITIVE LOGITS

 alebo

0.47

 ogrodow

0.47

 typelib

0.47

 zupeł

0.46

 singola

0.45

 Komunikasi

0.44

 individuale

0.44

or

0.44

 seduta

0.43

 creș

0.43

Activations Density 0.024%

phrases related to user rights and content moderation

remove or refuse

No Comments

No Known Activations