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

umo

-0.15

AndWait

-0.15

olib

-0.15

inery

-0.15

mb

-0.15

harma

-0.14

yo

-0.14

igure

-0.14

rique

-0.14

_approved

-0.14

POSITIVE LOGITS

acons

0.17

 third

0.16

 Hoover

0.15

opt

0.15

 hashed

0.15

 purposes

0.15

 Third

0.14

 interest

0.14

 anonymous

0.14

 unlawful

0.14

Activations Density 0.000%

No Known Activations

This feature has no known activations.