INDEX

Explanations

providing helpful information

np_max-act · gemini-2.0-flash

The neuron detects the model’s own self-referential framing language (e.g. “As an AI language model,” “I can…,” “However…”) that appears at the start of assistant responses.

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

^n

-0.06

 πρώ

-0.06

Coefficient

-0.06

 duyệt

-0.06

≤

-0.06

_SHADOW

-0.06

EZ

-0.06

ersions

-0.06

輕

-0.06

ansen

-0.06

POSITIVE LOGITS

깨

0.07

('<?

0.06

 redistribution

0.06

await

0.06

 accent

0.06

 Assassin

0.06

 trab

0.06

ुरक

0.06

alarını

0.06

equal

0.06

Activations Density 0.141%

providing helpful information

The neuron detects the model’s own self-referential framing language (e.g. “As an AI language model,” “I can…,” “However…”) that appears at the start of assistant responses.

No Comments

No Known Activations

providing helpful information

The neuron detects the model’s own self-referential framing language (e.g. “As an AI language model,” “I can…,” “However…”) that appears at the start of assistant responses.

No Comments

No Known Activations