INDEX

Explanations

numerical values or codes

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

Hexadecimal byte code sequences

np_acts-logits-general · gemini-2.0-flash

hex codes identifying programming structures

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

Referències

-0.52

principalColumn

-0.51

 aidant

-0.47

zv

-0.47

 suivre

-0.46

دانشنامهٔ

-0.45

кк

-0.44

kande

-0.43

脚注の使い方

-0.43

song

-0.43

POSITIVE LOGITS

Ngb

0.66

oredCriteria

0.63

ReusableCell

0.63

 Roumanie

0.62

addComponent

0.61

 AssemblyCulture

0.61

setVerticalGroup

0.60

TagHelper

0.59

Efq

0.57

 BrowserModule

0.56

Activations Density 0.372%

numerical values or codes

Hexadecimal byte code sequences

hex codes identifying programming structures

No Comments

No Known Activations