INDEX

Explanations

references to popular media and cultural content

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

mathematical expressions and errors

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

tagHelperRunner

-0.42

InjectAttribute

-0.39

存于互联网档案馆

-0.38

Hor

-0.36

orElse

-0.35

Blood

-0.34

teries

-0.34

 سرا

-0.34

Reading

-0.33

 fallu

-0.33

POSITIVE LOGITS

sizeCache

0.53

 المعيارى

0.44

 queſta

0.44

MenuGroup

0.44

 Meksiku

0.43

HandlerContext

0.43

تقاوى

0.43

mschrijving

0.41

 simplifié

0.41

 Roskov

0.41

Activations Density 0.004%

references to popular media and cultural content

mathematical expressions and errors

No Comments

No Known Activations