INDEX

Explanations

land areas and preserves

np_acts-logits-general · gemini-2.5-flash-lite

references to public lands, wildlife refuges, national parks, and protected natural areas.

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

The neuron flags numerical measurements of land or fire‐related scale—e.g. acres, hectares, percentages—often in the context of burned area, containment, or land-unit figures.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

************/

-1.17

 comesti

-1.10

 changeable

-1.09

PlayerPrefs

-1.08

ljiv

-1.08

存知

-1.05

 interminable

-1.05

rium

-1.05

лити

-1.03

'));

-1.02

POSITIVE LOGITS

it

1.66

 about

1.28

up

1.13

 that

1.10

 this

1.10

ászló

1.07

 mientras

1.06

 podczas

1.06

sopa

1.05

⇘

1.04

Activations Density 0.097%

land areas and preserves

references to public lands, wildlife refuges, national parks, and protected natural areas.

The neuron flags numerical measurements of land or fire‐related scale—e.g. acres, hectares, percentages—often in the context of burned area, containment, or land-unit figures.

No Comments

No Known Activations

land areas and preserves

references to public lands, wildlife refuges, national parks, and protected natural areas.

The neuron flags numerical measurements of land or fire‐related scale—e.g. acres, hectares, percentages—often in the context of burned area, containment, or land-unit figures.

No Comments

No Known Activations