INDEX

Explanations

reek

np_max-act · gemini-2.0-flash

The neuron strongly activates on LaTeX markup tokens (e.g. package names, document commands, math‐mode indicators), i.e. it detects LaTeX formatting.

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

详情

-0.06

uspendLayout

-0.06

.sha

-0.06

 semp

-0.06

ця

-0.06

 Leer

-0.06

頻

-0.06

(be

-0.06

ま

-0.06

lã

-0.06

POSITIVE LOGITS

reek

0.07

 KEEP

0.07

 singapore

0.07

 tweaking

0.06

ute

0.06

 uncertain

0.06

 quarterbacks

0.06

ensible

0.06

 registering

0.06

Activations Density 0.000%

reek

The neuron strongly activates on LaTeX markup tokens (e.g. package names, document commands, math‐mode indicators), i.e. it detects LaTeX formatting.

No Comments

No Known Activations