INDEX

Explanations

the term "deep" in various contexts, often related to depth or intensity in physical, emotional, or conceptual terms

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

deep X

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

 mList

-0.66

 laſſen

-0.61

 müſſen

-0.58

 getItemCount

-0.58

writeFieldEnd

-0.56

 MenuView

-0.56

 queſta

-0.56

.)}

-0.56



-0.55

 AssemblyCompany

-0.54

POSITIVE LOGITS

 deep

1.16

 Deep

1.14

 DEEP

1.12

Deep

1.10

deep

1.05

DEEP

0.93

 profundo

0.81

深

0.80

 deeply

0.79

 profundos

0.77

Activations Density 0.031%

the term "deep" in various contexts, often related to depth or intensity in physical, emotional, or conceptual terms

deep X

No Comments

No Known Activations