INDEX

Explanations

maintain

np_max-act · gemini-2.0-flash

This neuron specifically activates on the word “maintain” (and its close variants).

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 look

-0.06

Grupo

-0.06

 Choices

-0.06

 Scot

-0.06

 Quick

-0.06

 belong

-0.06

_clusters

-0.06

ックス

-0.06

 looks

-0.06

POSITIVE LOGITS

 maintained

0.11

 maintaining

0.11

 maintain

0.09

 maintains

0.09

Maint

0.08

 Maintain

0.08

mast

0.08

 Maint

0.08

 maintenance

0.08

Maintenance

0.08

Activations Density 0.025%

maintain

This neuron specifically activates on the word “maintain” (and its close variants).

No Comments

No Known Activations