INDEX

Explanations

code-related syntax, particularly related to variable declarations and assignments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 estekak

-0.69

miyorum

-0.50

AccessorTable

-0.49

ometra

-0.48

 PagesJaunes

-0.45

 fieldNum

-0.44

Xna

-0.43

󠁴

-0.41

GOTREF

-0.41

gestone

-0.40

POSITIVE LOGITS

SharedDtor

0.47

richTextPanel

0.41

fjspx

0.40

 poveznice

0.34

BASELINE

0.34

Intial

0.34

 awakeFromNib

0.34

 surla

0.33

 kasarigan

0.33

oucí

0.31

Activations Density 0.096%

code-related syntax, particularly related to variable declarations and assignments

No Comments

No Known Activations