INDEX

Explanations

numerical values presented in a fraction format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 3-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L3R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.3.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

-0.15

 Johnny

-0.14

FB

-0.14

 gains

-0.14

 tortured

-0.14

Yet

-0.14

ola

-0.13

Ãªn

-0.13

æĭ

-0.13

mu

-0.13

POSITIVE LOGITS

.infinity

0.16

ofday

0.16

.vaadin

0.16

reich

0.15

âĹıâĹı

0.15

Ð°Ð¹ÑĤ

0.15

_quant

0.15

iban

0.15

 Ø§ÙĦÙĨÙĩ

0.15

noinspection

0.15

Activations Density 0.010%

numerical values presented in a fraction format

No Comments

No Known Activations