INDEX

Explanations

phrases and terms related to foundational principles or frameworks for arguments or analyses

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

.Dispatch

-0.15

ijn

-0.15

iry

-0.14

 Kostenlose

-0.14

/pi

-0.14

 biases

-0.14

imd

-0.14

ips

-0.13

lsa

-0.13

 Preconditions

-0.13

POSITIVE LOGITS

yl

0.15

aqu

0.15

ÑĩÐµÐ¼

0.15

Ð³Ð¾Ð²

0.14

å¤®

0.14

chedulers

0.14

ampoo

0.14

ÙĬØ±Ø§ÙĨ

0.14

enti

0.14

cher

0.14

Activations Density 0.062%

phrases and terms related to foundational principles or frameworks for arguments or analyses

No Comments

No Known Activations