INDEX

Explanations

keywords related to various entities and classifications such as locations, subjects, and categories across multiple fields

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

animal, science, company, Android, Eagles

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

 ujednoznacz

-0.85

 utafitiHapana

-0.80

 nahilalakip

-0.73

 kasarigan

-0.66

Vidite

-0.65

 незавершена

-0.63

𝟸

-0.62

majánló

-0.59

 sexuelle

-0.59

səhifə

-0.58

POSITIVE LOGITS

setVerticalGroup

0.62

 <<<<<<<<<<<<<<

0.47

 CreateTagHelper

0.41

awtextra

0.31

③

0.31

AP

0.30

0.29

 <=",

0.29

StringCopy

0.29

oplayer

0.28

Activations Density 0.025%

keywords related to various entities and classifications such as locations, subjects, and categories across multiple fields

animal, science, company, Android, Eagles

No Comments

No Known Activations