INDEX

Explanations

No Explanations Found

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

/from

-0.20

odore

-0.16

uju

-0.15

 from

-0.14

ra

-0.14

 Nights

-0.14

feito

-0.13

ft

-0.13

erry

-0.13

POSITIVE LOGITS

mers

0.24

 experience

0.22

 left

0.21

 humble

0.21

mer

0.19

ager

0.18

Argb

0.18

now

0.18

 conception

0.17

 time

0.17

Activations Density 0.081%

No Comments

No Known Activations