INDEX

Explanations

Q\&A responses

np_max-act · gemini-2.0-flash

This neuron selectively activates on tokens in the model’s generated “assistant” responses (i.e. it flags words from the assistant’s output rather than user or system text).

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

콘

-0.07

 Committees

-0.07

 Born

-0.06

jerne

-0.06

폴

-0.06

OD

-0.06

'є

-0.06

 phạm

-0.06

داری

-0.06

 هاي

-0.06

POSITIVE LOGITS

.Interfaces

0.07

*>&

0.07

ほど

0.07

})}↵

0.07

OrFail

0.06

(visible

0.06

าตร

0.06

驶

0.06

=settings

0.06

_infos

0.06

Activations Density 0.131%

Q\&A responses

This neuron selectively activates on tokens in the model’s generated “assistant” responses (i.e. it flags words from the assistant’s output rather than user or system text).

No Comments

No Known Activations