INDEX

Explanations

through/via

np_max-act · gemini-2.0-flash

The neuron detects discourse markers that introduce mechanistic or causal links (e.g. “through,” “via,” “mechanistically”).

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

(geometry

-0.07

่าค

-0.07

vají

-0.06

ंब

-0.06

prak

-0.06

Btn

-0.06

_callbacks

-0.06

addons

-0.06

-circle

-0.06

 Analyst

-0.06

POSITIVE LOGITS

 code

0.07

 arch

0.07

 рад

0.06

 Georgetown

0.06

\''

0.06

via

0.06

^^^^

0.06

ochond

0.06

 sublime

0.06

Supply

0.06

Activations Density 0.019%

through/via

The neuron detects discourse markers that introduce mechanistic or causal links (e.g. “through,” “via,” “mechanistically”).

No Comments

No Known Activations