INDEX

Explanations

words related to user requests or inquiries for assistance

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

in the form

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

IndentedString

-0.63

 mergeFrom

-0.63

 utafitiHapana

-0.59

 saites

-0.56

@[+][

-0.56

تقاوى

-0.53

 ویکی‌پدی

-0.53

Jeografia

-0.52

webElementXpaths

-0.52

Personensuche

-0.51

POSITIVE LOGITS

WriteTagHelper

0.50

Joy

0.42

živ

0.42

rô

0.41

 Wear

0.41

Joy

0.41

 Grom

0.40

opr

0.40

Wear

0.40

 Bein

0.40

Activations Density 0.001%

words related to user requests or inquiries for assistance

in the form

No Comments

No Known Activations