INDEX

Explanations

will

np_max-act · gemini-2.0-flash

The neuron fires on occurrences of the verb “to be” (especially when used as an auxiliary in passive constructions, e.g. “is divided,” “will be started,” “is beautiful,” etc.).

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

_,↵

-0.07

bb

-0.06

ющее

-0.06

),

-0.06

_reports

-0.06

 convin

-0.06

 найбільш

-0.06

(reg

-0.06

怎

-0.06

battle

-0.06

POSITIVE LOGITS

_each

0.06

 Carolina

0.06

 Australian

0.06

 fortunes

0.06

getStyle

0.06

(numero

0.06

orian

0.06

 efficiencies

0.06

Alt

0.06

 Investor

0.06

Activations Density 0.356%

will

The neuron fires on occurrences of the verb “to be” (especially when used as an auxiliary in passive constructions, e.g. “is divided,” “will be started,” “is beautiful,” etc.).

No Comments

No Known Activations