INDEX

Explanations

references to people, positions, and factors within various contexts, indicating a focus on social dynamics and identities

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

Suppose, calculate, let

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

SourceChecksum

-0.54

astéroïdes

-0.50

Jeografia

-0.50

 Roskov

-0.48

NameInMap

-0.48

featureID

-0.46

hoeddwyd

-0.44

 Савезне

-0.44

.*")]

-0.42

 esternos

-0.41

POSITIVE LOGITS

 queſta

0.53

 auroit

0.50

 pouvoit

0.50

<unused79>

0.47

<unused8>

0.47

<unused41>

0.47

antaranya

0.47

<unused16>

0.47

<unused3>

0.47

<unused23>

0.47

Activations Density 0.096%

references to people, positions, and factors within various contexts, indicating a focus on social dynamics and identities

Suppose, calculate, let

No Comments

No Known Activations