INDEX

Explanations

references to contests or competitions

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

announcements of contest winners

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

بالإنجليزية

-0.49

 Ambiental

-0.49

 incorporar

-0.45

 שוליים

-0.45

 Marín

-0.45

Pb

-0.44

 Tiefen

-0.44

 quiera

-0.44

 materna

-0.44

Autowired

-0.43

POSITIVE LOGITS

 contest

1.21

 Contest

1.20

 CONTEST

1.13

Contest

1.09

 contests

1.05

contest

1.03

 competition

0.85

 contesting

0.83

 tournament

0.83

 concurso

0.82

Activations Density 0.002%

references to contests or competitions

announcements of contest winners

No Comments

No Known Activations