INDEX

Explanations

references to code snippets and data types

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-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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 flecha

-0.47

 produjo

-0.44

windowFixed

-0.44

 InputDecoration

-0.43

stylers

-0.43

gsmål

-0.42

 prioridad

-0.39

 producido

-0.39

 causado

-0.39

 obtenido

-0.38

POSITIVE LOGITS

 cref

0.84

sizeof

0.70

 instanceof

0.66

 AssemblyProduct

0.62

memberof

0.62

 sizeof

0.62

WEBPACK

0.54

__(/*!

0.53

 nameof

0.53

QMetaType

0.52

Activations Density 0.117%

references to code snippets and data types

No Comments

No Known Activations