INDEX

Explanations

negations and expressions of uncertainty or exceptions in statements

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

when followed by 'the' or 'why'

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

 الحره

-0.43

 Roskov

-0.43

ftagPool

-0.38

 disambiguazione

-0.37

قایناق‌لار

-0.37

delwed

-0.36

AntiForgeryToken

-0.36

 Italijani

-0.35

WebServlet

-0.34

Спасылкі

-0.34

POSITIVE LOGITS

󠁢

0.51

 autorytatywna

0.50

GraphicsUnit

0.49

terase

0.47

MENAFN

0.46

 unknownFields

0.45

ULTS

0.44

 мәкал

0.44

 sogar

0.44

しかも

0.43

Activations Density 0.111%

negations and expressions of uncertainty or exceptions in statements

when followed by 'the' or 'why'

No Comments

No Known Activations