INDEX

Explanations

error

np_max-act · gemini-2.0-flash

The neuron spots instances of the “plain‐error” review discussion in the opinion text, including its accompanying numeric error‐value annotations.

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

 residences

-0.07

 ConsoleColor

-0.06

 pole

-0.06

“Well

-0.06

fh

-0.06

ţi

-0.06

 Sofia

-0.06

 termination

-0.06

esi

-0.06

-helper

-0.06

POSITIVE LOGITS

 instructional

0.07

Ner

0.07

 сы

0.06

presso

0.06

竜

0.06

periment

0.06

村

0.06

jectives

0.06

 Oliveira

0.06

 Від

0.06

Activations Density 0.000%

error

The neuron spots instances of the “plain‐error” review discussion in the opinion text, including its accompanying numeric error‐value annotations.

No Comments

No Known Activations