INDEX

Explanations

specific names and references related to a show or characters

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

Michael Arbuthnot Australia

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

 without

-0.29

imp

-0.29

Toi

-0.28

❹

-0.28

瑯

-0.28

 صوتيه

-0.27

nezeu

-0.27

 يتيمه

-0.27

nloa

-0.27

 regular

-0.26

POSITIVE LOGITS

 ExecuteAsync

0.66

linawan

0.63

 queſta

0.63

 AssemblyCompany

0.63

 Geiſt

0.62

ðsíða

0.61

########.

0.60

 increí

0.59

ロウィン

0.59

 indígen

0.58

Activations Density 0.000%

specific names and references related to a show or characters

Michael Arbuthnot Australia

No Comments

No Known Activations