INDEX

Explanations

specific measurements and properties of nodes and edges in a graph or algorithmic context

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

AnimationsModule

-0.47

RTLI

-0.46

 للمعارف

-0.43

PostExecute

-0.43

ImageIO

-0.39

 समीक्षक

-0.39

setof

-0.37

AndroidJUnit

-0.36

copg

-0.36

⇑

-0.36

POSITIVE LOGITS

 gynhyrchwyd

0.43

 escritorio

0.43

værende

0.41

embatan

0.40

 calculées

0.40

 mecán

0.40

ligner

0.39

 Polskiej

0.38

 Савезне

0.38

 Comprometido

0.38

Activations Density 0.235%

specific measurements and properties of nodes and edges in a graph or algorithmic context

No Comments

No Known Activations