INDEX

Explanations

references to architecture

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

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

zier

-0.22

ocking

-0.17

ovich

-0.16

orious

-0.16

oje

-0.16

ocha

-0.15

áº©y

-0.15

omorphic

-0.15

Ñİ

-0.15

oded

-0.15

POSITIVE LOGITS

ipel

0.35

etypes

0.32

bishop

0.30

iving

0.29

itect

0.29

angel

0.28

ivist

0.28

etype

0.26

ived

0.26

aic

0.24

Activations Density 0.010%

references to architecture

No Comments

No Known Activations