INDEX

Explanations

repeated conjunctions and logical qualifiers in arguments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 jadx

-0.38

 quadro

-0.36

).

-0.32

 нови

-0.31

 Dorado

-0.31

xrTableCell

-0.31

+#+#

-0.31

newblock

-0.30

#+#

-0.30

 cobre

-0.30

POSITIVE LOGITS

лтамалар

0.69

 nahilalakip

0.66

DebuggerNonUser

0.58

تقاوى

0.58

 كومونز

0.56

ISupport

0.56

Šaltiniai

0.55

 ویکی‌پدی

0.54

 autorytatywna

0.54

Personensuche

0.54

Activations Density 0.632%

repeated conjunctions and logical qualifiers in arguments

No Comments

No Known Activations