INDEX

Explanations

phrases indicating expectations or predictions regarding outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

-FIRST

-0.14

wat

-0.12

urette

-0.12

Ð½Ð¸ÑĤ

-0.12

analyze

-0.12

vise

-0.12

çĸ²

-0.12

trys

-0.12

ynchronize

-0.12

atta

-0.12

POSITIVE LOGITS

 grace

0.26

 headline

0.23

top

0.20

 duke

0.19

 land

0.19

 rank

0.19

 rival

0.19

 rule

0.18

 outs

0.18

 command

0.18

Activations Density 0.207%

phrases indicating expectations or predictions regarding outcomes

No Comments

No Known Activations