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

apk

-0.20

.hr

-0.18

idge

-0.16

åĳ½

-0.16

odge

-0.15

ange

-0.15

ace

-0.15

icator

-0.15

deÅŁ

-0.14

otto

-0.14

POSITIVE LOGITS

azer

0.19

erence

0.19

aru

0.18

adian

0.17

RD

0.16

iken

0.16

enci

0.16

ISCO

0.16

inz

0.15

cube

0.15

Activations Density 0.034%

No Comments

No Known Activations