INDEX

Explanations

award-related terms associated with films

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

vician

-0.47

工程师

-0.42

 vaksinasi

-0.41

 voisins

-0.40

timmt

-0.40

 dovol

-0.39

 razonable

-0.39

zbęd

-0.38

 méthodes

-0.38

 sottile

-0.38

POSITIVE LOGITS

yntaxException

0.51

endregion

0.47

archiviato

0.44

 Film

0.43

 Schell

0.42

bolista

0.42

 international

0.42

VPB

0.42

 juri

0.41

 Савезне

0.41

Activations Density 0.011%

award-related terms associated with films

No Comments

No Known Activations