INDEX

Explanations

references to individuals with "former" titles or positions

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

former roles and entities

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

IntoConstraints

-0.77

RenderAtEndOf

-0.71

 estekak

-0.64

 snippetHide

-0.63

randomUUID

-0.60

tagHelperRunner

-0.58

出版年

-0.57

 متعلقه

-0.56

SequentialGroup

-0.56

ServletException

-0.54

POSITIVE LOGITS

 former

0.66

Former

0.66

 Former

0.59

former

0.55

 mantan

0.54

 býval

0.52

 ehemalige

0.51

 ehemaligen

0.41

ex

0.40

 ehemal

0.40

Activations Density 0.018%

references to individuals with "former" titles or positions

former roles and entities

No Comments

No Known Activations