INDEX

Explanations

references to formal discussions, meetings, and arguments

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

oral argument

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

heça

-0.50

 resourceCulture

-0.46

 spalle

-0.45

 noiva

-0.45

AndEndTag

-0.45

HtmlAttribute

-0.45

 dinding

-0.44

RenderAtEndOf

-0.43

 ContentValues

-0.43

 wireType

-0.43

POSITIVE LOGITS

المكان

0.46

/**

0.39

󠁴

0.38

μβρίου

0.38

Calla

0.38

 OGSÅ

0.36

()?;

0.36

irl

0.36

flashdata

0.36

relsen

0.35

Activations Density 0.065%

references to formal discussions, meetings, and arguments

oral argument

No Comments

No Known Activations