INDEX

Explanations

phrases that indicate successful functionality or effectiveness

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

functioning

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

unhofer

-0.41

oneta

-0.41

arition

-0.40

fieldset

-0.39

 Sixty

-0.39

annah

-0.38

{/

-0.36

 Caution

-0.36

fread

-0.36

Captor

-0.35

POSITIVE LOGITS

Worked

0.79

 Worked

0.77

worked

0.72

 fonctionner

0.70

 worked

0.70

 WORKING

0.68

 funcionar

0.68

 ExecuteAsync

0.66

 Works

0.66

 WORKS

0.66

Activations Density 0.035%

phrases that indicate successful functionality or effectiveness

functioning

No Comments

No Known Activations