INDEX

Explanations

code

np_max-act · gemini-2.0-flash

The neuron activates on tokens that form hexadecimal identifier fragments (e.g., parts of URLs or commit hashes in hex).

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

subtotal

-0.07

(mi

-0.07

quared

-0.06

products

-0.06

overview

-0.06

_qty

-0.06

 바이

-0.06

Sup

-0.06

uffman

-0.06

POSITIVE LOGITS

 अलग

0.07

 fract

0.07

 Ок

0.07

 xnxx

0.07

 χαρα

0.07

 läng

0.07

 ổn

0.06

XE

0.06

 yanlış

0.06

 Contr

0.06

Activations Density 0.035%

code

The neuron activates on tokens that form hexadecimal identifier fragments (e.g., parts of URLs or commit hashes in hex).

No Comments

No Known Activations