INDEX

Explanations

question marks

np_max-act · gemini-2.0-flash

signals of user questions in chat logs, especially conversational boundary tokens and question-form cues like punctuation and interrogative structure.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AQ

-0.07

 NAMES

-0.06

bursement

-0.06

く

-0.06

Suc

-0.06

 Entire

-0.06

 Proble

-0.06

OCK

-0.06

 Trades

-0.06

 всего

-0.06

POSITIVE LOGITS

 stationed

0.07

 tiles

0.06

pair

0.06

emonic

0.06

 ayır

0.06

 huge

0.06

ประส

0.06

šení

0.06

 pumpkin

0.06

realDonaldTrump

0.06

Activations Density 0.165%

question marks

signals of user questions in chat logs, especially conversational boundary tokens and question-form cues like punctuation and interrogative structure.

No Comments

No Known Activations