INDEX

Explanations

Code/punctuation

np_max-act · gemini-2.0-flash

This neuron activates on tokens that are part of example demonstrations or code/math notation (e.g. braces, punctuation, format specifiers, header markers) rather than on ordinary running text.

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

TTY

-0.06

ЎыџN

-0.06

\\.

-0.06

 Salt

-0.06

Salt

-0.06

거래가

-0.06

UTERS

-0.06

.request

-0.06

 Nazis

-0.06

 sunday

-0.05

POSITIVE LOGITS

(bot

0.07

Jer

0.07

 '',
↵

0.06

 чит

0.06

 competit

0.06

 approximation

0.06

 likelihood

0.06

сли

0.06

.ISupportInitialize

0.06

�

0.06

Activations Density 0.072%

Code/punctuation

This neuron activates on tokens that are part of example demonstrations or code/math notation (e.g. braces, punctuation, format specifiers, header markers) rather than on ordinary running text.

No Comments

No Known Activations