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

-orange

-0.21

arta

-0.18

agers

-0.18

-0.17

itchens

-0.16

arah

-0.15

ively

-0.15

orders

-0.15

zer

-0.14

acker

-0.14

POSITIVE LOGITS

ignal

0.22

iginal

0.22

IENTATION

0.21

tega

0.20

ogonal

0.20

ifold

0.19

amental

0.19

ourke

0.19

IGIN

0.18

acular

0.18

Activations Density 0.075%

No Comments

No Known Activations