INDEX

Explanations

instances of strong emphasis or extreme importance

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

academic/scientific citations

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

 houſe

-0.42

DataAnnotations

-0.41

Trayectoria

-0.41

 ſch

-0.40

dataclass

-0.38

getMinutes

-0.38

 ſte

-0.37

 ſta

-0.36

 tranſ

-0.36

 ſol

-0.36

POSITIVE LOGITS

 Италијани

0.75

 nakalista

0.69

 Numerade

0.60

 EconPapers

0.59

<eos>

0.58

 Савезне

0.54

OGND

0.54

 ostavi

0.52

EndGlobal

0.49

رشف

0.48

Activations Density 0.002%

instances of strong emphasis or extreme importance

academic/scientific citations

No Comments

No Known Activations