INDEX

Explanations

terms related to calculations and their processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

fall

-0.17

rms

-0.16

olt

-0.16

HEST

-0.16

IMARY

-0.15

lund

-0.15

apers

-0.14

eler

-0.14

mary

-0.14

POSITIVE LOGITS

/cal

0.17

irim

0.16

ypse

0.15

evice

0.15

ifornia

0.15

mlink

0.15

tras

0.14

ãĥ³ãĥĦ

0.14

ador

0.14

uppe

0.14

Activations Density 0.028%

terms related to calculations and their processes

No Comments

No Known Activations