INDEX

Explanations

phrases or expressions of gratitude and appreciation

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

programming context, code Snippets

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

 ſch

-0.64

 itſelf

-0.62

 myſelf

-0.60

 zuſammen

-0.60

 Geſch

-0.58

 beſch

-0.58

 Wikimedijinoj

-0.58

 geſch

-0.58

ésultats

-0.58

iſchen

-0.57

POSITIVE LOGITS

tagHelperRunner

0.35

oneofs

0.34

Karriere

0.33

 escapar

0.33

OK

0.32

<?

0.31

…………………………………………

0.31

estr

0.30

axel

0.30

imageshack

0.30

Activations Density 0.190%

phrases or expressions of gratitude and appreciation

programming context, code Snippets

No Comments

No Known Activations