INDEX

Explanations

specific names and terms related to various domains, including codes, historical figures, and measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 gagne

-0.46

 sanitaires

-0.45

 scolaire

-0.43

 vectorielles

-0.43

 alimentaires

-0.42

 beraber

-0.40

 verschieden

-0.40

appliquent

-0.40

 avoit

-0.40

 sauvages

-0.39

POSITIVE LOGITS

AsUp

0.94

Datuak

0.92

StoryboardSegue

0.82

 GenerationType

0.81

TintMode

0.79

 MainAxisSize

0.74

StreetMap

0.73

PreferredItem

0.73

ValueStyle

0.73

NamedQueries

0.73

Activations Density 0.757%

specific names and terms related to various domains, including codes, historical figures, and measurements

No Comments

No Known Activations