INDEX

Explanations

references to numerical values and color codes, particularly in programming contexts

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

transQ

-0.68

IBOutlet

-0.56

 Infór

-0.56

 CallOverrides

-0.53

 queſto

-0.52

 noDo

-0.51

TagMode

-0.51

 erſten

-0.51

ſelben

-0.51

 متعلقه

-0.51

POSITIVE LOGITS

 appunto

0.42

页面存档备份

0.36

raltar

0.35

 sollen

0.34

発

0.33

ьому

0.30

 statt

0.29

 invece

0.29

 instead

0.29

 mittel

0.29

Activations Density 0.005%

references to numerical values and color codes, particularly in programming contexts

No Comments

No Known Activations