INDEX

Explanations

usernames

np_max-act · gemini-2.0-flash

The neuron primarily detects the placeholder name tokens (e.g. “NAME_1,” “NAME_2,” etc.) used for referring to characters.

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

Laf

-0.07

 Team

-0.07

 dialog

-0.07

 hitter

-0.06

 logos

-0.06

ΟΓ

-0.06

oro

-0.06

dog

-0.06

resse

-0.06

Truth

-0.06

POSITIVE LOGITS

        ↵        ↵        ↵

0.07

②

0.06

%%%

0.06

ście

0.06

量

0.06

클

0.06

 convince

0.06

 detain

0.06

[['

0.06

Activations Density 0.026%

usernames

The neuron primarily detects the placeholder name tokens (e.g. “NAME_1,” “NAME_2,” etc.) used for referring to characters.

No Comments

No Known Activations