INDEX

Explanations

brand names and company names referenced in the text

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

company names and associations

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

 sija

-0.43

expandindo

-0.40

Lesen

-0.39

sidemargin

-0.39

 terça

-0.38

 okuyayım

-0.36

ChildScrollView

-0.35

 jotka

-0.35

transQ

-0.33

 betweenstory

-0.33

POSITIVE LOGITS

ſelf

0.62

ulemon

0.61

dflare

0.59

 Majefty

0.58

 ResponseEntity

0.56

barco

0.54

 ✭✭

0.53

 SAMSUNG

0.52

 Chriftian

0.52

antd

0.52

Activations Density 0.019%

brand names and company names referenced in the text

company names and associations

No Comments

No Known Activations