INDEX

Explanations

specific identifiers or terms that indicate important elements, possibly in contexts like programming, documentation, or legal matters

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Sauer

-0.50

 Raub

-0.49

 sauer

-0.48

 blasphemy

-0.47

 betweenstory

-0.46

 Calvo

-0.44

 sorcery

-0.44

 erratic

-0.44

 Zwiebel

-0.43

 EDITORS

-0.43

POSITIVE LOGITS

BibitemShut

0.65

bibfield

0.51

новништво

0.48

 avoient

0.46

ScopeManager

0.45

viewWillAppear

0.44

 päät

0.44

nologies

0.43

AnchorStyles

0.43

 læng

0.42

Activations Density 0.022%

specific identifiers or terms that indicate important elements, possibly in contexts like programming, documentation, or legal matters

No Comments

No Known Activations