INDEX

Explanations

terms related to attention and tracking details in various contexts

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

pay attention to, keep track of

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 houver

-0.42

 supuestamente

-0.42

 neby

-0.41

 berke

-0.39

 tanggung

-0.38

买的

-0.38

 collègues

-0.38

があれば

-0.37

 Veuillez

-0.37

 nincs

-0.36

POSITIVE LOGITS

ImageContext

0.83

featureID

0.82

存于互联网档案馆

0.69

 JpaRepository

0.67

WriteTagHelper

0.66

arraycopy

0.65

setVerticalGroup

0.65

EndInit

0.63

 مشين

0.62

躇

0.62

Activations Density 0.024%

terms related to attention and tracking details in various contexts

pay attention to, keep track of

No Comments

No Known Activations