INDEX

Explanations

pl

np_max-act · gemini-2.0-flash

The neuron fires on the leading subword chunks of complex technical terms—i.e. Greek/Latin prefixes such as “mult-,” “pl-” (plasma-), “plur-,” “supra-,” “orth-,” “ter-,” 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

brain

-0.07

 tongues

-0.07

课

-0.07

_player

-0.07

 Plains

-0.07

 ром

-0.07

bands

-0.06

败

-0.06

-avatar

-0.06

frey

-0.06

POSITIVE LOGITS

 Undo

0.07

 Null

0.07

(return

0.07

{Name

0.06

//----------------------------------------------------------------------------------------------------------------

0.06

org

0.06

(animated

0.06

 opts

0.06

(ValueError

0.06

 ++)↵

0.06

Activations Density 0.681%

pl

The neuron fires on the leading subword chunks of complex technical terms—i.e. Greek/Latin prefixes such as “mult-,” “pl-” (plasma-), “plur-,” “supra-,” “orth-,” “ter-,” etc.

No Comments

No Known Activations

pl

The neuron fires on the leading subword chunks of complex technical terms—i.e. Greek/Latin prefixes such as “mult-,” “pl-” (plasma-), “plur-,” “supra-,” “orth-,” “ter-,” etc.

No Comments

No Known Activations