INDEX

Explanations

references to negation or conditions in statements

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

gridy

-0.59

 geſch

-0.59

jsii

-0.57



-0.56

rawn

-0.55

 propOrder

-0.55

 المعيارى

-0.54

 ſei

-0.54

 ſind

-0.53

 createStore

-0.53

POSITIVE LOGITS

 proof

0.31

WarningLevel

0.30

 BoxFit

0.29

 tromper

0.28

 يتيمه

0.28

//!

0.28

enschappelijke

0.28

epresidente

0.28

 juger

0.27

ä

0.27

Activations Density 0.092%

references to negation or conditions in statements

No Comments

No Known Activations