INDEX

Explanations

instances of the word "and" in various forms and contexts

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

phrases with "and" or "/" separators

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

 queima

-0.38

houn

-0.38

 Aurora

-0.38

cap

-0.38

 Grove

-0.37

 Logic

-0.37

arian

-0.37

tabl

-0.36

 typelib

-0.36

bro

-0.36

POSITIVE LOGITS

Personendaten

0.61

 للاسماء

0.56

 utafitiHapana

0.48

évaluateur

0.47

KindOfClass

0.47

 initComponents

0.45

__*/

0.45

pioneer

0.43

Tembelea

0.43

ResumeLayout

0.42

Activations Density 0.081%

instances of the word "and" in various forms and contexts

phrases with "and" or "/" separators

No Comments

No Known Activations