INDEX

Explanations

programming-related constructs, particularly functions and their definitions

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 typelib

-0.53

 zvuk

-0.52

 miniaturka

-0.51

 kapture

-0.49



-0.47

enderror

-0.46

 strijd

-0.44

 wohin

-0.44

 Sunak

-0.44

 zvuky

-0.43

POSITIVE LOGITS

ArgsConstructor

0.88

UserScript

0.63

 '\\;'

0.62

 ReactDOM

0.46

classmethod

0.45

//

0.43

ISupport

0.42

ⓧ

0.40

RefNanny

0.39

 الحره

0.39

Activations Density 0.139%

programming-related constructs, particularly functions and their definitions

No Comments

No Known Activations