INDEX

Explanations

Negation/Argumentation

np_max-act · gemini-2.0-flash

This neuron activates on modal and negation words (e.g. should, can’t, not, possible) that express ability or permission.

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

-category

-0.07

trainer

-0.07

 partir

-0.07

อาหาร

-0.06

.Tasks

-0.06

 Latina

-0.06

 tốc

-0.06

уб

-0.06

 Holl

-0.06

伟

-0.06

POSITIVE LOGITS

as

0.06

 özellik

0.06

_Items

0.06

_WRITE

0.06

_effects

0.06

nost

0.06

.the

0.06

.SizeMode

0.06

 Declarations

0.05

:H

0.05

Activations Density 0.177%

Negation/Argumentation

This neuron activates on modal and negation words (e.g. should, can’t, not, possible) that express ability or permission.

No Comments

No Known Activations