INDEX

Explanations

identifiers related to code and programming

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

setVerticalGroup

-0.64

原始内容存档于

-0.56

didSet

-0.52

 MainAxisSize

-0.51

styleable

-0.50

IntoConstraints

-0.49

#+#

-0.49

pexpr

-0.47

 Jefus

-0.47

 跳转至

-0.46

POSITIVE LOGITS

astéroïdes

0.45

 Wikispecies

0.35

 extract

0.35

<eos>

0.35

Biographie

0.34

 Brau

0.32

 referenties

0.31

cohol

0.31

 adjust

0.31

 Bluffs

0.30

Activations Density 0.074%

identifiers related to code and programming

No Comments

No Known Activations