INDEX

Explanations

code

np_max-act · gemini-2.0-flash

The neuron strongly activates on shell‐style punctuation tokens—especially backslashes used for line continuations and “=” in environment‐variable or flag assignments.

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

 pigeon

-0.07

-cert

-0.07

Alarm

-0.07

 alarms

-0.06

 oper

-0.06

 SENSOR

-0.06

 celebrities

-0.06

 tails

-0.06

ция

-0.06

 turtles

-0.06

POSITIVE LOGITS

EMPTY

0.07

购买

0.06

 JAXB

0.06

 χρόνια

0.06

.В

0.06

.B

0.06

 qualifying

0.06

 şekilde

0.06

Activations Density 0.015%

code

The neuron strongly activates on shell‐style punctuation tokens—especially backslashes used for line continuations and “=” in environment‐variable or flag assignments.

No Comments

No Known Activations