INDEX

Explanations

contextual references to ongoing actions and their implications

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

gameplay interesting

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

ReusableCell

-0.43

 ineffective

-0.36

Seiten

-0.35

 Schell

-0.33

 adept

-0.33

 Silver

-0.33

 jambes

-0.33

SourceChecksum

-0.32

 Heller

-0.31

be

-0.30

POSITIVE LOGITS

LookAnd

0.82

دانشنامهٔ

0.74

fillType

0.70

findpost

0.68

 يتيمه

0.66

Бахар

0.63

RenderAtEndOf

0.60

 متعلقه

0.60

 noDo

0.59

 kasarigan

0.59

Activations Density 0.026%

contextual references to ongoing actions and their implications

gameplay interesting

No Comments

No Known Activations