INDEX

Explanations

references to research methodologies and tools

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

Follows "which," "you," or a symbol

np_acts-logits-general · gemini-2.0-flash

revised guidelines

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-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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits



-0.41

 Alone

-0.40

OGND

-0.39

zab

-0.38

 CanadaChoose

-0.37

Gaz

-0.37

@[+][

-0.36

laient

-0.35

Alone

-0.35

mag

-0.34

POSITIVE LOGITS

ConstraintMaker

0.55

rungsseite

0.55

⟬

0.49

الحياه

0.49

endpush

0.48

ImageContext

0.47

PMailer

0.46

 springfox

0.45

AndEndTag

0.45

queryset

0.45

Activations Density 0.645%

references to research methodologies and tools

Follows "which," "you," or a symbol

revised guidelines

No Comments

No Known Activations