INDEX

Explanations

references to attention or focus in 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 @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

rices

-0.18

inton

-0.17

hack

-0.17

hma

-0.16

itude

-0.16

acker

-0.16

imuth

-0.16

pora

-0.15

Injector

-0.15

hood

-0.15

POSITIVE LOGITS

ested

0.27

estation

0.24

uned

0.23

ests

0.23

esting

0.23

itud

0.21

en

0.21

aining

0.21

0.20

0.19

Activations Density 0.006%

references to attention or focus in various contexts

No Comments

No Known Activations