INDEX

Explanations

terms related to growth or increase in scale

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

fully

-0.17

Æł

-0.16

anners

-0.16

ialized

-0.16

lessly

-0.16

zelf

-0.15

erman

-0.15

plash

-0.15

utow

-0.15

ourney

-0.15

POSITIVE LOGITS

 upon

0.27

 Upon

0.21

 into

0.19

Upon

0.18

/import

0.17

hor

0.17

able

0.15

ary

0.15

/general

0.15

avier

0.15

Activations Density 0.033%

terms related to growth or increase in scale

No Comments

No Known Activations