INDEX

Explanations

People descriptions

np_max-act · gemini-2.0-flash

The neuron detects brief appositive or relative clauses that attribute a person’s credentials, role, or affiliation (e.g. “who…,” “historian,” “works for,” 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

 хот

-0.07

豊

-0.07

 кораб

-0.07

Jac

-0.07

 आकर

-0.06

decor

-0.06

ушки

-0.06

-Qaeda

-0.06

地区

-0.06

/stats

-0.06

POSITIVE LOGITS

 COUNT

0.07

lickr

0.06

lim

0.06

触

0.06

 lipid

0.06

 мають

0.06

 communicate

0.06

 Timothy

0.06

ltra

0.06

imin

0.06

Activations Density 0.037%

People descriptions

The neuron detects brief appositive or relative clauses that attribute a person’s credentials, role, or affiliation (e.g. “who…,” “historian,” “works for,” etc.).

No Comments

No Known Activations

People descriptions

The neuron detects brief appositive or relative clauses that attribute a person’s credentials, role, or affiliation (e.g. “who…,” “historian,” “works for,” etc.).

No Comments

No Known Activations