INDEX

Explanations

with

np_max-act · gemini-2.0-flash

specific terms or technologies that are not commonly recognized or require clarification.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

The neuron detects the model’s expressions of apology and unfamiliarity (e.g. “I’m sorry, but I’m not familiar with…”).

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

.pres

-0.07

 equipped

-0.06

 Hawth

-0.06

 fract

-0.06

िछल

-0.06

 Kosovo

-0.06

(commands

-0.06

ıyor

-0.06

pH

-0.06

 wholesome

-0.06

POSITIVE LOGITS

jur

0.07

 prez

0.07

 Components

0.06

/pub

0.06

plib

0.06

hus

0.06

 لل

0.06

 lumber

0.06

 الملك

0.06

 heal

0.06

Activations Density 0.033%

with

specific terms or technologies that are not commonly recognized or require clarification.

The neuron detects the model’s expressions of apology and unfamiliarity (e.g. “I’m sorry, but I’m not familiar with…”).

No Comments

No Known Activations