INDEX

Explanations

possibilities

np_max-act · gemini-2.0-flash

The neuron fires on modal‐perfect constructions (phrases like “could have done,” “could have been,” etc.) indicating hypothetical or counterfactual actions.

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

amples

-0.07

DK

-0.07

.rf

-0.07

_primary

-0.06

ählt

-0.06

vf

-0.06

ameron

-0.06

.media

-0.06

 headquartered

-0.06

POSITIVE LOGITS

 could

0.07

 seas

0.06

 disadv

0.06

 Trab

0.06

 spontaneously

0.06

 інт

0.06

 streets

0.06

 moment

0.06

/)

0.06

/trans

0.06

Activations Density 0.024%

possibilities

The neuron fires on modal‐perfect constructions (phrases like “could have done,” “could have been,” etc.) indicating hypothetical or counterfactual actions.

No Comments

No Known Activations