INDEX

Explanations

terms related to competitive events and rankings

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

competition stages

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

 Италијани

-0.75

 disambiguazione

-0.73

 surla

-0.63

SourceChecksum

-0.61

Portail

-0.59

 Italijanski

-0.58

@"/

-0.56



-0.56

 Infórmanos

-0.56

 ReactDOM

-0.56

POSITIVE LOGITS

 round

0.41

 preliminary

0.40

 playoffs

0.38

 elimination

0.38

 ration

0.37

 Elimination

0.36

 compartment

0.35

 مظ

0.35

 Qualifying

0.35

 rounds

0.34

Activations Density 0.015%

terms related to competitive events and rankings

competition stages

No Comments

No Known Activations