INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.16.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

hetto

-0.16

ols

-0.15

rets

-0.15

æĿ

-0.14

pheric

-0.14

uja

-0.14

amd

-0.13

 Implements

-0.13

gow

-0.13

inq

-0.13

POSITIVE LOGITS

/gpl

0.16

ERG

0.16

áŀ¶

0.15

orz

0.15

ç·Ĵ

0.15

yte

0.15

BuilderFactory

0.15

 overdue

0.14

ationToken

0.14

ako

0.14

Activations Density 0.000%

No Known Activations

This feature has no known activations.