INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

agus

-0.16

 getattr

-0.15

.plugins

-0.15

quier

-0.14

UY

-0.14

 masked

-0.14

manuel

-0.14

 ÙħÙĤØ§Ùħ

-0.13

 Aston

-0.13

alty

-0.13

POSITIVE LOGITS

CWE

0.32

sink

0.22

Sink

0.21

_sink

0.21

bad

0.20

 sink

0.20

 Sink

0.20

 sinks

0.19

Bad

0.18

 INCIDENTAL

0.17

Activations Density 0.000%

No Known Activations

This feature has no known activations.