INDEX

Explanations

words related to the beginning stages or introductions of various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

ess

-0.19

alim

-0.17

pire

-0.17

annes

-0.16

ongan

-0.16

ensch

-0.16

alen

-0.15

rey

-0.15

al

-0.15

Ø§ÙĨÙĩ

-0.15

POSITIVE LOGITS

mente

0.27

izers

0.25

isation

0.25

-stage

0.24

iser

0.23

isms

0.23

 stages

0.23

ising

0.22

izations

0.20

ling

0.20

Activations Density 0.017%

words related to the beginning stages or introductions of various contexts

No Comments

No Known Activations