INDEX

Explanations

articles and determiner phrases

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

a noun

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

 GenerationType

-0.68

 препратки

-0.62

تقاوى

-0.59

bootstrapcdn

-0.59

 Infórmanos

-0.58

 PeEnEo

-0.55

 kaarangay

-0.55

󠁣

-0.53

 otomatig

-0.53

ViewFeatures

-0.53

POSITIVE LOGITS

bcryptjs

0.42

0.40

Jam

0.40

Set

0.40

Jam

0.39

↵↵

0.39

Web

0.35

FailureListener

0.34

fra

0.34

toja

0.34

Activations Density 0.088%

articles and determiner phrases

a noun

No Comments

No Known Activations