INDEX

Explanations

personal thoughts/opinions

np_max-act · gemini-2.0-flash

The neuron detects authorial self‐references combined with emphatic or intensifying adverbs (e.g. “I really…,” “I can’t believe…,” “I think…”).

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

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

PP

-0.07

amin

-0.07

 wall

-0.06

HELL

-0.06

ion

-0.06

*$

-0.06

DIR

-0.06

}','

-0.06

 ucwords

-0.06

payments

-0.06

POSITIVE LOGITS

licit

0.07

основ

0.07

 جل

0.06

asyarak

0.06

zenie

0.06

 กล

0.06

-slider

0.06

_alg

0.06

 Pradesh

0.06

 케이

0.06

Activations Density 0.077%

personal thoughts/opinions

The neuron detects authorial self‐references combined with emphatic or intensifying adverbs (e.g. “I really…,” “I can’t believe…,” “I think…”).

No Comments

No Known Activations