INDEX

Explanations

references to hills or elevated land formations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

-0.16

sak

-0.15

baum

-0.15

obra

-0.14

ooke

-0.14

yne

-0.14

roker

-0.14

Ã¼re

-0.14

vais

-0.14

zure

-0.14

POSITIVE LOGITS

side

0.47

iard

0.40

top

0.39

crest

0.30

arious

0.28

ides

0.27

ier

0.24

endale

0.23

-top

0.22

erman

0.21

Activations Density 0.012%

references to hills or elevated land formations

No Comments

No Known Activations