INDEX

Explanations

references to blog posts and helpful resources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

abd

-0.17

-0.15

.UserInfo

-0.14

anes

-0.14

à¥Ģà¤²

-0.14

ossa

-0.13

_bulk

-0.13

ANA

-0.13

literal

-0.13

Ã¤Ã¤

-0.13

POSITIVE LOGITS

oldt

0.16

ogui

0.16

immel

0.15

 published

0.14

ittings

0.14

ipple

0.14

osyal

0.14

ordion

0.14

-archive

0.14

é¹¿

0.14

Activations Density 0.087%

references to blog posts and helpful resources

No Comments

No Known Activations