INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 vá»¥

-0.17

kili

-0.15

 vulnerable

-0.14

ãĥ§

-0.14

 preventative

-0.14

aurus

-0.14

åį«çĶŁ

-0.13

.Trace

-0.13

_sensitive

-0.13

ationToken

-0.13

POSITIVE LOGITS

 diet

0.28

 exercises

0.26

 dietary

0.25

 exercise

0.25

 supplements

0.24

 Diet

0.22

 diets

0.22

 Exercises

0.22

 medication

0.22

 lifestyle

0.22

Activations Density 0.301%

No Comments

No Known Activations