INDEX

Explanations

numerical addresses and codes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

æĤ

-0.15

ennon

-0.14

 Sting

-0.14

ainer

-0.14

-0.13

ocale

-0.13

rows

-0.13

obby

-0.13

 roman

-0.13

POSITIVE LOGITS

.cx

0.16

anz

0.15

olle

0.15

agrid

0.15

vX

0.15

lett

0.14

erli

0.14

icos

0.14

wyn

0.14

ÃĹ↵↵

0.14

Activations Density 0.006%

numerical addresses and codes

No Comments

No Known Activations