INDEX

Explanations

proper nouns and references to individuals

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

former elected retired opened testified

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

 ordini

-0.29

 varsa

-0.28

 lleno

-0.28

rů

-0.28

 bezeichneter

-0.27

 grossen

-0.27

 siglos

-0.26

jenigen

-0.26

 secoli

-0.25

LookAnd

-0.25

POSITIVE LOGITS

LabelTagHelper

0.66

 Paglinawan

0.64

RectangleBorder

0.63

parsedMessage

0.63

 مرئيه

0.63

tvguidetime

0.63

msgTypes

0.62

ſehen

0.60

 ſei

0.60

ſelben

0.59

Activations Density 0.016%

proper nouns and references to individuals

former elected retired opened testified

No Comments

No Known Activations