INDEX

Explanations

technical or programming-related terms and structures

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

archaic spellings and turn separators

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

Географиясе

-0.41

webElementXpaths

-0.39

SerializedName

-0.36

 conformité

-0.35

<eos>

-0.34

belajaran

-0.33

kmäler

-0.32

オリティ

-0.31

 tecnici

-0.31

 دان

-0.31

POSITIVE LOGITS

 CreateTagHelper

0.81

 Houſe

0.66

 houſe

0.66

 greateſt

0.65

Datuak

0.64

 leaſt

0.63

 myſelf

0.62

Efq

0.60

 purpoſe

0.59

 ſte

0.58

Activations Density 0.006%

technical or programming-related terms and structures

archaic spellings and turn separators

No Comments

No Known Activations