INDEX

Explanations

references to hierarchical structures or tree-like arrangements in data

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

structured data tags and lists

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 snippetHide

-0.52

umented

-0.48

 intptr

-0.47

enablog

-0.47

قایناقلار

-0.46

abetes

-0.46

risse

-0.46

 kaynağından

-0.46

matology

-0.45

hyrchwyd

-0.45

POSITIVE LOGITS

 Frage

0.38

 שם

0.37

UnknownFieldSet

0.37

EndContext

0.37

 túnel

0.36

 Isaacs

0.35

 escalera

0.34

つか

0.34

 gelatina

0.34

Diweddarwch

0.33

Activations Density 0.080%

references to hierarchical structures or tree-like arrangements in data

structured data tags and lists

No Comments

No Known Activations