INDEX

Explanations

:

np_max-act · gemini-2.0-flash

This neuron detects the introductory tokens that mark the start of an answer (e.g., “A:”, “Answer:”, or similar lead-in words).

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

The start of an answer or response section (tokens marking the beginning of an answer).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

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

.FragmentManager

-0.06

version

-0.06

-labelledby

-0.06

ोप

-0.06

gst

-0.06

osal

-0.06

yt

-0.06

 پاد

-0.06

 harvested

-0.06

 treatment

-0.06

POSITIVE LOGITS

 Από

0.07

 Auswahl

0.07

 fills

0.07

�

0.07

 mundo

0.06

除了

0.06

�

0.06

ค

0.06

Polygon

0.06

 mondo

0.06

Activations Density 0.169%

:

This neuron detects the introductory tokens that mark the start of an answer (e.g., “A:”, “Answer:”, or similar lead-in words).

The start of an answer or response section (tokens marking the beginning of an answer).

No Comments

No Known Activations