INDEX

Explanations

words and phrases indicating progress, advancement, or elevation in status or position

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

utow

-0.15

kir

-0.15

Sto

-0.14

ikel

-0.14

esModule

-0.13

 Rudy

-0.13

 escorte

-0.13

 heading

-0.13

lob

-0.13

amber

-0.13

POSITIVE LOGITS

 ranks

0.29

 rank

0.25

 rankings

0.21

èĩ³

0.21

rank

0.21

_rank

0.20

 into

0.20

Rank

0.20

.rank

0.19

 ladder

0.18

Activations Density 0.042%

words and phrases indicating progress, advancement, or elevation in status or position

No Comments

No Known Activations