INDEX

Explanations

elements related to programming logic and data handling

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

New Auto-Interp

Configuration

fnlp/Llama3_1-8B-Base-LXM-8x/Llama3_1-8B-Base-L22M-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.hook_mlp_out

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

bard

-0.15

igram

-0.14

TOTYPE

-0.14

 DropIndex

-0.14

stab

-0.14

Jer

-0.14

 Gill

-0.14

ä¸Ī

-0.13

orado

-0.13

POSITIVE LOGITS

 Increment

0.53

increment

0.51

 increment

0.51

Increment

0.48

 increments

0.46

 Incre

0.46

 decrement

0.43

 incr

0.43

.increment

0.42

 increase

0.41

Activations Density 0.276%

elements related to programming logic and data handling

No Comments

No Known Activations