INDEX

Explanations

references to spatial relationships or directions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 2-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

 fevere

-0.57

 ſch

-0.54

 ſche

-0.53

 pleaſure

-0.53

 otomatig

-0.51

 ſtate

-0.51

 InputDecoration

-0.50

 Normdatei

-0.50

 productId

-0.49

 للمعارف

-0.49

POSITIVE LOGITS

 along

1.07

along

0.90

Along

0.88

 Along

0.85

 ALONG

0.82

 вдоль

0.62

 entlang

0.58

 junto

0.54

沿

0.54

 alongside

0.53

Activations Density 0.151%

references to spatial relationships or directions

No Comments

No Known Activations