INDEX

Explanations

references to placeholder pages

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

enton

-0.17

ansa

-0.17

ah

-0.16

Kit

-0.15

anj

-0.15

 nonce

-0.15

IOD

-0.14

erif

-0.14

iod

-0.14

onyms

-0.14

POSITIVE LOGITS

'gc

0.18

ëł¹

0.15

ÏģÎ¿Ïħ

0.15

AIT

0.14

Fullscreen

0.14

ypse

0.14

ĥn

0.14

edor

0.13

à¸±à¸ģà¸ģ

0.13

Ð¾ÑģÑĤÐµÐ¹

0.13

Activations Density 0.000%

references to placeholder pages

No Comments

No Known Activations