INDEX

Explanations

I

np_max-act · gemini-2.0-flash

The neuron broadly responds to high-frequency, common English tokens—i.e. function words (like “I,” “it,” “for,” “and,” “to”) and punctuation—rather than rare content words.

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

first-person, self-referential statements describing personal actions, plans, or experiences (typical of forum-style posts).

oai_token-act-pair · gpt-5 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

 gusto

-0.07

 stitching

-0.07

，你

-0.07

_layout

-0.07

Watch

-0.07

.Ş

-0.06

ircraft

-0.06

aning

-0.06

<Component

-0.06

utr

-0.06

POSITIVE LOGITS

ریه

0.08

ніч

0.07

MOT

0.06

 eski

0.06

//$

0.06

-pol

0.06

_TMP

0.06

vo

0.06

Crear

0.06

 grandi

0.06

Activations Density 0.091%

I

The neuron broadly responds to high-frequency, common English tokens—i.e. function words (like “I,” “it,” “for,” “and,” “to”) and punctuation—rather than rare content words.

first-person, self-referential statements describing personal actions, plans, or experiences (typical of forum-style posts).

No Comments

No Known Activations

I

The neuron broadly responds to high-frequency, common English tokens—i.e. function words (like “I,” “it,” “for,” “and,” “to”) and punctuation—rather than rare content words.

first-person, self-referential statements describing personal actions, plans, or experiences (typical of forum-style posts).

No Comments

No Known Activations