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

ruba

-0.17

anke

-0.16

imitives

-0.16

LEM

-0.16

udem

-0.15

VML

-0.15

ulta

-0.14

CSC

-0.14

adata

-0.14

zed

-0.13

POSITIVE LOGITS

AIS

0.35

ais

0.26

ais

0.25

 sucking

0.23

 suck

0.22

 sucked

0.22

 choir

0.21

cho

0.20

 rehearsal

0.19

AI

0.19

Activations Density 0.000%

No Known Activations

This feature has no known activations.