INDEX

Explanations

structured elements or categorization in various 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

EndInit

-0.60

mbic

-0.43

RegistryLite

-0.38

✨:

-0.36

edonia

-0.35

 Core

-0.35

ResumeLayout

-0.34

findpost

-0.34

BaseAdapter

-0.34

 core

-0.33

POSITIVE LOGITS

 linkovi

0.61

 հղումներ

0.58

Autoritní

0.57

تقاوى

0.52

лтамалар

0.51

wapV

0.49

ьажоргаш

0.47

الحياه

0.47

EnableWeb

0.45

 Administrativna

0.44

Activations Density 0.090%

structured elements or categorization in various contexts

No Comments

No Known Activations