INDEX

Explanations

references to specific types and attributes of objects

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

Followed by a specifier or detail

np_acts-logits-general · gemini-2.0-flash

specify a category

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

bolista

-0.37

setopt

-0.35

 beziehen

-0.34

 fakt

-0.34

PostBack

-0.32

 Nursing

-0.32

mmol

-0.32

 bezieht

-0.32

logrus

-0.31

ifstream

-0.31

POSITIVE LOGITS

 defaultstate

0.73

richTextPanel

0.58

featureID

0.57

 kind

0.56

toMatchSnapshot

0.56

 Certain

0.54

ftagPool

0.54

 CERTAIN

0.52

StoryboardSegue

0.51

特定

0.51

Activations Density 0.576%

references to specific types and attributes of objects

Followed by a specifier or detail

specify a category

No Comments

No Known Activations