INDEX

Explanations

numerical information and statistics

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

ÙĦÙĤ

-0.15

 Bios

-0.15

 Alley

-0.14

&R

-0.14

allery

-0.14

emsp

-0.14

assa

-0.14

 Bris

-0.13

essel

-0.13

&C

-0.13

POSITIVE LOGITS

crement

0.15

ahat

0.14

xCD

0.14

shima

0.14

erh

0.14

Ã¼Ã§Ã¼k

0.14

mud

0.14

_mk

0.13

_OC

0.13

CREMENT

0.13

Activations Density 0.004%

numerical information and statistics

No Comments

No Known Activations