INDEX

Explanations

words and phrases associated with lists and formatting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_16k/average_l0_77

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.17.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

 Monfieur

-1.28

Efq

-1.23

 CHAPITRE

-1.23

 Anſ

-1.21

 Diſ

-1.18

 pleaſure

-1.14

 Jefus

-1.12

 ainfi

-1.11

 Houſe

-1.11

 Majefty

-1.10

POSITIVE LOGITS

чо

0.58

le

0.54

ja

0.49

kon

0.47

 Bloc

0.47

pol

0.47

les

0.46

ze

0.46

dan

0.46

Activations Density 1.156%

words and phrases associated with lists and formatting

No Comments

No Known Activations