INDEX

Explanations

Question/problem introduction

np_max-act · gemini-2.0-flash

The neuron detects sentence‐initial directive or instruction markers (e.g. “Put,” “Don’t,” “Let,” “If,” “Each,” “We”) that introduce user guidelines or rules.

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

immers

-0.07

 suç

-0.07

 invoked

-0.07

ился

-0.06

ті

-0.06

 міг

-0.06

 Cheers

-0.06

Kelly

-0.06

Roger

-0.06

UC

-0.06

POSITIVE LOGITS

 espacio

0.07

 แพ

0.07

 seper

0.06

 Coordinates

0.06

Sec

0.06

 mouseY

0.06

_atom

0.06

 radiant

0.06

=a

0.06

검

0.06

Activations Density 0.039%

Question/problem introduction

The neuron detects sentence‐initial directive or instruction markers (e.g. “Put,” “Don’t,” “Let,” “If,” “Each,” “We”) that introduce user guidelines or rules.

No Comments

No Known Activations

Question/problem introduction

The neuron detects sentence‐initial directive or instruction markers (e.g. “Put,” “Don’t,” “Let,” “If,” “Each,” “We”) that introduce user guidelines or rules.

No Comments

No Known Activations