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

elon

-0.15

abant

-0.14

arasÄ±

-0.14

eron

-0.14

-Col

-0.14

.synthetic

-0.14

ERRU

-0.14

æ§

-0.14

 CACHE

-0.14

 prest

-0.14

POSITIVE LOGITS

bad

0.17

anken

0.16

Bad

0.16

BAD

0.15

 Fallen

0.15

man

0.15

.man

0.15

 shared

0.14

 Bair

0.14

 mann

0.14

Activations Density 0.000%

No Known Activations

This feature has no known activations.