INDEX

Explanations

elements related to positional information in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 letters

-0.26

 Strecke

-0.24

LETTERS

-0.24

 drewni

-0.23

betrieb

-0.23

лист

-0.23

 Letters

-0.23

 gồm

-0.23

Tôi

-0.23

 male

-0.22

POSITIVE LOGITS

IntoConstraints

0.91

 위치

0.82

 Position

0.81

sidemargin

0.80

 ubicación

0.78

 location

0.76

makeConstraints

0.75

addPreferredGap

0.75

 position

0.74

########.

0.74

Activations Density 0.952%

elements related to positional information in various contexts

No Comments

No Known Activations