INDEX

Explanations

references to presence and impact in various contexts

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

all or everywhere

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

 inter

-0.38

__((

-0.35

 عدم

-0.32

createServer

-0.32

 being

-0.32

 tissu

-0.31

Og

-0.31

ahon

-0.31

MP

-0.31

 inclined

-0.31

POSITIVE LOGITS

 every

0.87

 everywhere

0.82

 ovunque

0.81

 seluruh

0.77

 EVERY

0.73

everywhere

0.73

every

0.73

 invokingState

0.69

 wszyst

0.68

 semua

0.67

Activations Density 0.061%

references to presence and impact in various contexts

all or everywhere

No Comments

No Known Activations