INDEX

Explanations

proper nouns, particularly names and brands

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

New Auto-Interp

Configuration

Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post

Features

65,536

Data Type

float32

Hook Name

blocks.25.hook_resid_post

Hook Layer

Architecture

gated

Context Size

1,024

Dataset

Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ARAM

-0.17

ãĤ·ãĤ¢

-0.14

ÐµÐ²ÐµÑĢ

-0.14

omain

-0.14

<->

-0.14

 Unused

-0.14

Ð¸Ð·Ð½ÐµÑģ

-0.14

ienda

-0.14

 Disqus

-0.13

æ¢°

-0.13

POSITIVE LOGITS

II

0.33

III

0.29

II

0.24

 project

0.24

-II

0.24

IV

0.23

ii

0.21

III

0.20

project

0.19

 Project

0.19

Activations Density 0.033%

proper nouns, particularly names and brands

No Comments

No Known Activations