INDEX

Explanations

information related to gameplay mechanics and strategies in a specific online game

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

making the

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

ſelf

-0.69

ſelves

-0.62

 purpoſe

-0.61

 faſt

-0.59

 tranſ

-0.58

 anſ

-0.58

 fubject

-0.57

 ſtate

-0.56

 ſever

-0.56

 ſta

-0.56

POSITIVE LOGITS

 Fulda

0.44

 Persson

0.43

 underrated

0.40

 NSCoder

0.37

 glistening

0.37

 STRATEGY

0.37

 StatelessWidget

0.37

 POSITION

0.36

rvats

0.36

 famed

0.36

Activations Density 0.020%

information related to gameplay mechanics and strategies in a specific online game

making the

No Comments

No Known Activations