INDEX

Explanations

language related to changes and transformations

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

increase or decrease

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

 propOrder

-0.60

zzleHttp

-0.60

enderror

-0.53

󠁴

-0.50

jspx

-0.49

 referrerpolicy

-0.45

 surla

-0.43

 indisponible

-0.42

 zugesch

-0.42

 Compulsory

-0.42

POSITIVE LOGITS

 مشين

0.47

inição

0.41

 globalization

0.38

decrease

0.36

ckså

0.36

increase

0.35

 contrário

0.35

 contradiction

0.34

Diwedd

0.34

 purpoſe

0.34

Activations Density 0.092%

language related to changes and transformations

increase or decrease

No Comments

No Known Activations