INDEX

Explanations

references to user interface components and states in a programming context

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

page or user contexts

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

WriteAttribute

-0.54

 виправивши

-0.49

पया

-0.47

amssymb

-0.43

 postIndex

-0.43

tagHelperRunner

-0.43

SerializedName

-0.42

nonumber

-0.41

 PyTuple

-0.41

uxxxx

-0.41

POSITIVE LOGITS

 HomeScreen

0.69

 surla

0.61

IntoConstraints

0.54

 חיצוניים

0.53

 Obrador

0.50

SceneManagement

0.46

 laboratoire

0.43

Rhestr

0.42

 الرياضيه

0.42

󠁢

0.42

Activations Density 0.005%

references to user interface components and states in a programming context

page or user contexts

No Comments

No Known Activations