INDEX

Explanations

phrases that discuss potential actions and improvements

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

how to improve or take action

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

EndContext

-0.53

PMailer

-0.51

новниш

-0.51

iNdEx

-0.47

principalTable

-0.46

 iconst

-0.45

libft

-0.44

__(/*!

-0.43

rippa

-0.43

!*\

-0.43

POSITIVE LOGITS

 enhance

0.42

enhance

0.41

 WAYS

0.40

/**

0.39

 enhancement

0.39

IntoConstraints

0.39

 künftig

0.38

ViewFeatures

0.37

 ways

0.37

NegativeButton

0.37

Activations Density 0.024%

phrases that discuss potential actions and improvements

how to improve or take action

No Comments

No Known Activations