INDEX

Explanations

programming and coding terminology related to functions and objects

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

C++ code syntax

np_acts-logits-general · gemini-2.0-flash

code structure parsing

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

kana

-0.42

 Photocase

-0.41

ckets

-0.41

cookieParser

-0.40

 &___

-0.40

extAlignment

-0.39

GINIA

-0.39

 scream

-0.39

 invade

-0.38

Concept

-0.38

POSITIVE LOGITS

MLLoader

0.66

Personendaten

0.63

 defaultstate

0.48

StoryboardSegue

0.44

 fusca

0.44

 enfans

0.44

Geplaatst

0.43

 lateinit

0.42

UnsafeEnabled

0.41

BeginContext

0.40

Activations Density 1.033%

programming and coding terminology related to functions and objects

C++ code syntax

code structure parsing

No Comments

No Known Activations