INDEX

Explanations

patterns in numerical data and sequences

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

Date formats with delimiters

np_acts-logits-general · gemini-2.0-flash

date separators like '/' and '-'

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

ckså

-0.46

laştır

-0.45

 Inscrivez

-0.45

ագրություններ

-0.45

 églises

-0.44

 themſelves

-0.43

 couvercle

-0.42

 leaſt

-0.42

 touristes

-0.42

ſelf

-0.41

POSITIVE LOGITS

Nationalité

0.51

fromLTRB

0.49

GEBURTSDATUM

0.49

subsubsection

0.48

 transfieras

0.48

Beijos

0.47

+#+

0.46

 Roskov

0.46

FromNib

0.45

+"/"+

0.45

Activations Density 0.205%

patterns in numerical data and sequences

Date formats with delimiters

date separators like '/' and '-'

No Comments

No Known Activations