INDEX

Explanations

Programming code

np_max-act · gemini-2.0-flash

This neuron activates on programming language primitive type names (like string, number, null, undefined) and related type‐annotation tokens.

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

 observers

-0.07

Fuse

-0.07

CellValue

-0.07

postcode

-0.06

-carousel

-0.06

_surface

-0.06

经过

-0.06

_fac

-0.06

unik

-0.06

Brains

-0.06

POSITIVE LOGITS

 розмі

0.07

 BOOL

0.07

_BOOL

0.07

 teď

0.07

 레이

0.06

AND

0.06

 CENT

0.06

ALAR

0.06

sto

0.06

 seç

0.06

Activations Density 0.028%

Programming code

This neuron activates on programming language primitive type names (like string, number, null, undefined) and related type‐annotation tokens.

No Comments

No Known Activations