INDEX

Explanations

terms related to various domains such as contracts, market research, and psychological concepts

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

suffixes and conceptual words

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

+#+

-0.72

 EconPapers

-0.70

 propOrder

-0.58

porters

-0.56

Skocz

-0.54

zwi

-0.54

ducers

-0.53

itarian

-0.53

moid

-0.52

⧠

-0.52

POSITIVE LOGITS

Derbyniad

0.41

awtextra

0.40

Controllo

0.38

 autorytatywna

0.37

WebElementEntity

0.37

 Мексичка

0.34

كويكب

0.33

 overall

0.33

 мәкал

0.31

IUrlHelper

0.31

Activations Density 0.113%

terms related to various domains such as contracts, market research, and psychological concepts

suffixes and conceptual words

No Comments

No Known Activations