INDEX

Explanations

terms related to the concept of "partial."

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

ulner

-0.17

usement

-0.16

format

-0.15

enberg

-0.15

irma

-0.15

hammer

-0.15

anders

-0.14

asca

-0.14

icken

-0.14

edBy

-0.14

POSITIVE LOGITS

/full

0.33

ity

0.29

ities

0.22

ake

0.22

led

0.20

ledge

0.19

OOK

0.19

ties

0.19

AKE

0.19

Argb

0.18

Activations Density 0.016%

terms related to the concept of "partial."

No Comments

No Known Activations