INDEX

Explanations

references to situations and contexts in the text

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

preceding a comma

np_acts-logits-general · gemini-2.0-flash

phrases indicating purpose or consequence

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 geſch

-0.74

 ſei

-0.71

 Normdatei

-0.70

 kaarangay

-0.70

 ſondern

-0.68

SequentialGroup

-0.68

 müſſen

-0.68

 queſta

-0.67

 verſch

-0.67

 ſeine

-0.67

POSITIVE LOGITS

 also

0.36

is

0.34

0.33

 includes

0.32

ins

0.30

 implies

0.28

 effectively

0.27

 means

0.27

 allows

0.25

Activations Density 0.487%

references to situations and contexts in the text

preceding a comma

phrases indicating purpose or consequence

No Comments

No Known Activations