INDEX

Explanations

phrases indicating parts of a whole or subsets within a context

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.32

 double

-0.28

OnAdd

-0.28

-0.26

pod

-0.25

ta

-0.25

 Byers

-0.24

 tack

-0.24

-0.23

POSITIVE LOGITS

 ویکی‌پدیا

0.86

 autorytatywna

0.85

 متعلقه

0.77

rungsseite

0.77

ſſung

0.77

 utafitiHapana

0.76

 otomatig

0.75

principalColumn

0.75



0.73

 Infórmanos

0.72

Activations Density 0.090%

phrases indicating parts of a whole or subsets within a context

No Comments

No Known Activations