INDEX

Explanations

temporal references and numerical information

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

punctuation followed by a word

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

 ujednoznacz

-0.47

 disponibilités

-0.44

adpleegd

-0.42

OptionsMenu

-0.42

 milho

-0.40

bootstrapcdn

-0.39

 huelga

-0.38

prüche

-0.37

 okno

-0.37

 strike

-0.37

POSITIVE LOGITS

balleur

0.58

ToScroll

0.50

 كومونز

0.49

Founded

0.46

RTLR

0.45

 للمعارف

0.44

cipl

0.44

を経て

0.43

 Founded

0.43

slidesPer

0.43

Activations Density 0.024%

temporal references and numerical information

punctuation followed by a word

No Comments

No Known Activations