INDEX

Explanations

phrases indicating completion or finishing an action

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

finished or done with

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

 scolaire

-0.42

rások

-0.40

 africain

-0.40

Debido

-0.38

 scuole

-0.37

 increase

-0.37

 anuales

-0.36

 acela

-0.36

 fréquence

-0.36

POSITIVE LOGITS

DONE

0.74

 Done

0.73

Done

0.72

isDone

0.71

 DONE

0.70

contentLoaded

0.69

 Finished

0.68

 propOrder

0.66

elesai

0.66

 متعلقه

0.62

Activations Density 0.012%

phrases indicating completion or finishing an action

finished or done with

No Comments

No Known Activations