INDEX

Explanations

modal verbs and their conjugations in various languages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_7/width_16k/average_l0_92

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.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

ſelf

-0.77

ſelves

-0.71

 plufieurs

-0.63

 myſelf

-0.63

 noDo

-0.63

 faſt

-0.62

ArrowToggle

-0.62

 ſte

-0.62

 affari

-0.62

 dezelve

-0.61

POSITIVE LOGITS

 wird

1.31

 werden

1.20

 wordt

1.09

 wurde

1.08

wird

1.03

 Wird

0.99

 worden

0.98

 werd

0.95

Wird

0.93

 wurden

0.91

Activations Density 0.119%

modal verbs and their conjugations in various languages

No Comments

No Known Activations