INDEX

Explanations

references to variables and features in a dataset

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 AssemblyCulture

-0.64

windowFixed

-0.55

 незавершена

-0.54

AddTagHelper

-0.53

Geplaatst

-0.53

 للاسماء

-0.52

Gambas

-0.51

 Мексичка

-0.50

UnsafeEnabled

-0.49

 ujednoznacz

-0.49

POSITIVE LOGITS

ChildScrollView

0.54

 Keuangan

0.40

info

0.40

Requires

0.37

 расход

0.36

 Finanz

0.35

ijoje

0.35

 emissions

0.34

ường

0.34

 inputs

0.34

Activations Density 0.011%

references to variables and features in a dataset

No Comments

No Known Activations