INDEX

Explanations

phrases that indicate processes or actions aimed at achieving specific outcomes

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

order to achieve

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

UserScript

-0.59

ⓧ

-0.59

IndentedString

-0.52

BagConstraints

-0.48

bootstrapcdn

-0.47

 externi

-0.45

istoitu

-0.45

PreferredItem

-0.42

 propOrder

-0.42

MetaObject

-0.42

POSITIVE LOGITS

才能

0.59

 <<<<<<<<<<<<<<

0.56

 properly

0.49

properly

0.47

なければ

0.45

しなければ

0.45

才會

0.44

才可以

0.43

survival

0.41

才会

0.41

Activations Density 0.044%

phrases that indicate processes or actions aimed at achieving specific outcomes

order to achieve

No Comments

No Known Activations