INDEX

Explanations

references to individuals or groups, specifically using the word "who."

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

those who

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

NewUrlParser

-0.34

Condol

-0.32

tvguidetime

-0.30

+-

-0.28

 кӀ

-0.27

帖最后由

-0.27

又是一

-0.26

but

-0.26

,...

-0.26

 heißt

-0.25

POSITIVE LOGITS

ðsíða

0.73

 Italij

0.68

 ddelwed

0.67

iſchen

0.67

 Houſe

0.66

ۗ

0.66

ésultats

0.66

<unused43>

0.65

<unused8>

0.65

[@BOS@]

0.65

Activations Density 0.040%

references to individuals or groups, specifically using the word "who."

those who

No Comments

No Known Activations