INDEX

Explanations

pathfinding algorithms and code

np_max-act · gemini-2.0-flash

The neuron is triggered by the high-level description of a program’s goal or “intention” statement (the summary sentence that explains what the code is meant to do).

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Downtown

-0.07

morph

-0.07

zij

-0.06

 Alta

-0.06

 Bite

-0.06

 bonding

-0.06

UpDown

-0.06

Lyn

-0.06

 granularity

-0.06

ونی

-0.06

POSITIVE LOGITS

 enabled

0.07

 image

0.06

 Object

0.06

.obtain

0.06

 uveden

0.06

 آز

0.06

ชนะ

0.06

 unint

0.06

 Assuming

0.06

FFECT

0.06

Activations Density 0.015%

pathfinding algorithms and code

The neuron is triggered by the high-level description of a program’s goal or “intention” statement (the summary sentence that explains what the code is meant to do).

No Comments

No Known Activations