INDEX

Explanations

numerical values or mentions of quantities

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

publications and dates

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-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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 bowiem

-0.42

⎠

-0.38

 totiž

-0.36

تفصیلات

-0.35

 löytyy

-0.35

 wystarczy

-0.34

 metade

-0.33

 vastaan

-0.33

 przecież

-0.33

 táctil

-0.33

POSITIVE LOGITS

RegressionTest

0.55

rungsseite

0.52

发表于

0.50

GetKeyDown

0.50

printStackTrace

0.50

########.

0.49

homonymie

0.48

 Falla

0.47

esca

0.47

 שוליים

0.46

Activations Density 0.051%

numerical values or mentions of quantities

publications and dates

No Comments

No Known Activations