INDEX

Explanations

variations of the prefix "un-"

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

 è¶

-0.18

Ð°ÑĤÐ¸Ð²

-0.15

 Hawth

-0.15

aver

-0.15

ÙĬØ¹

-0.14

igi

-0.14

chner

-0.14

 ÑħÑĥÐ´

-0.14

 Lower

-0.14

CKER

-0.13

POSITIVE LOGITS

sworth

0.16

shelf

0.16

OLON

0.15

vail

0.15

izen

0.15

åį

0.15

 buckle

0.15

pto

0.14

 buck

0.14

Ø³ÙĨ

0.14

Activations Density 0.021%

variations of the prefix "un-"

No Comments

No Known Activations