INDEX

Explanations

phrases indicating causal relationships or conditions with the word "for."

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

jee

-0.17

ilm

-0.16

ively

-0.15

ties

-0.15

emes

-0.14

for

-0.14

atable

-0.14

å¥Ī

-0.14

erosis

-0.14

POSITIVE LOGITS

-profit

0.29

 sake

0.27

bidden

0.27

geries

0.26

ays

0.25

aging

0.24

 instance

0.23

 purposes

0.22

feit

0.21

asm

0.21

Activations Density 0.712%

phrases indicating causal relationships or conditions with the word "for."

No Comments

No Known Activations