INDEX

Explanations

words related to performance, scoring, and competitive outcomes

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

lead

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

AndEndTag

-0.76

 ModelExpression

-0.66

expandindo

-0.62

 disambiguazione

-0.58

httphttps

-0.55

nasse

-0.54

 AssemblyCompany

-0.54

Chham

-0.53

 betweenstory

-0.53

endregion

-0.52

POSITIVE LOGITS

 chặt

0.31

 insurmountable

0.30

ModelSerializer

0.28

<<

0.27

 Hillsborough

0.27

inspace

0.27

offsets

0.27

 abgesch

0.26

 Pils

0.26

スタン

0.26

Activations Density 0.010%

words related to performance, scoring, and competitive outcomes

lead

No Comments

No Known Activations