INDEX

Explanations

programming-related terms or keywords

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

tonsoft

-0.53

 WASTE

-0.49

 LINES

-0.48

 STONE

-0.47

KindOfClass

-0.47

 LABORATOR

-0.46

 DISTANCE

-0.45

far

-0.45

 VOLTAGE

-0.45

 DAMAGE

-0.45

POSITIVE LOGITS

Or

0.64

 évêque

0.56

 jsPsych

0.54

And

0.53

Re

0.53

Us

0.51

 wikipagina

0.51

De

0.49

Pr

0.49

Co

0.49

Activations Density 2.208%

programming-related terms or keywords

No Comments

No Known Activations