INDEX

Explanations

key topics related to categories, events, and individuals associated with specific actions or states

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

crime, victory, or info

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

 veille

-0.41

tênis

-0.40

 agujas

-0.39

 manguera

-0.39

 ganchillo

-0.38

 nubes

-0.37

 botellas

-0.37

 enfermera

-0.36

 aguja

-0.35

POSITIVE LOGITS

 nakalista

0.71

 HttpNotFound

0.65

RunWith

0.62

 ویکی‌پدی

0.60

الإنجليزية

0.59

providedIn

0.57

OGND

0.54

HasAnnotation

0.51

 springfox

0.51

principalColumn

0.50

Activations Density 0.010%

key topics related to categories, events, and individuals associated with specific actions or states

crime, victory, or info

No Comments

No Known Activations