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

ehler

-0.20

 ourselves

-0.16

olin

-0.15

/basic

-0.15

 discriminator

-0.14

 neighboring

-0.14

ube

-0.14

avor

-0.14

æĪĳä»¬

-0.14

 unfavorable

-0.14

POSITIVE LOGITS

 Tanner

0.25

 fucking

0.23

 fuck

0.22

 fucked

0.21

 fucks

0.21

 Fuck

0.18

fuck

0.18

 Fucking

0.17

 cunt

0.17

Fuck

0.16

Activations Density 0.000%

No Known Activations

This feature has no known activations.