INDEX

Explanations

structured references to items or sections, particularly in a numbered format

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

躇

-0.52

ագրություններ

-0.51

WriteLiteral

-0.50

scriptcase

-0.48

enumi

-0.47

 ddelweddau

-0.47



-0.46

 يتيمه

-0.43

@@@@@

-0.42

getWriter

-0.41

POSITIVE LOGITS

RotationOrder

0.41

reportWebVitals

0.41

instancetype

0.37

ReusableCell

0.36

 Signalez

0.36

uncu

0.35

rinfo

0.35

 Vorbe

0.35

 defaultstate

0.35

finement

0.34

Activations Density 0.530%

structured references to items or sections, particularly in a numbered format

No Comments

No Known Activations