INDEX

Explanations

the conjunction "and."

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

ninger

-0.21

oje

-0.18

edl

-0.16

ED

-0.16

edics

-0.16

ington

-0.15

leo

-0.15

chn

-0.15

mediate

-0.15

IEWS

-0.15

POSITIVE LOGITS

olph

0.28

eur

0.24

ele

0.23

eed

0.23

een

0.22

orf

0.22

ale

0.21

ean

0.21

ee

0.21

eb

0.21

Activations Density 0.040%

the conjunction "and."

No Comments

No Known Activations