INDEX

Explanations

references to significant numerical values or statistics in scientific discussions

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

starts with figure or citation

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

MigrationBuilder

-0.65

UrlResolution

-0.56

 Biôgrafia

-0.56

Билгалдахарш

-0.54

PullParser

-0.53

 Italijani

-0.53

RegistryLite

-0.51

󠁮

-0.51

 utafitiHapana

-0.50

 defaultstate

-0.50

POSITIVE LOGITS

↵↵

0.41

↵↵↵↵

0.40

 дописавши

0.39

 Martín

0.37

 transforma

0.36

Hauptartikel

0.36

 Martin

0.36

 placa

0.35

Queued

0.35

 Exactos

0.34

Activations Density 0.009%

references to significant numerical values or statistics in scientific discussions

starts with figure or citation

No Comments

No Known Activations