INDEX

Explanations

references to specific languages

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

language names

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

ThroughAttribute

-0.65

expandindo

-0.60

TargetException

-0.59

makeConstraints

-0.58

وفاته

-0.55

lifted

-0.55

 Autorizaciones

-0.54

(!__

-0.54

tvguidetime

-0.54

✨:

-0.54

POSITIVE LOGITS

 Arabic

0.60

 Hindi

0.59

 language

0.57

 English

0.56

 Language

0.56

Hindi

0.55

 Urdu

0.52

Arabic

0.51

 languages

0.51

 German

0.50

Activations Density 0.008%

references to specific languages

language names

No Comments

No Known Activations