INDEX

Explanations

specific identifiers or codes that might relate to names or key subjects

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

repeated short sequences

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

 Normdatei

-0.78

tagHelperRunner

-0.77

 propOrder

-0.74

󠁣

-0.73

 мәкал

-0.73

 Geſch

-0.71

MLLoader

-0.67

存于互联网档案馆

-0.67

 Geiſt

-0.66

arangay

-0.65

POSITIVE LOGITS

↵↵

0.47

↵

0.44

0.42

<strong>

0.41

0.39

0.38

...

0.37

 Egger

0.37

0.36

↵↵↵↵

0.36

Activations Density 0.068%

specific identifiers or codes that might relate to names or key subjects

repeated short sequences

No Comments

No Known Activations