INDEX

Explanations

pronouns and verbs that indicate user actions and instructions

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

icot

-0.19

Ã§uk

-0.15

ãĥ¼ãĤº

-0.15

 knowledge

-0.15

izza

-0.15

enzyme

-0.14

 Laure

-0.14

ĥ½

-0.14

pur

-0.14

 IsNot

-0.14

POSITIVE LOGITS

can

0.15

dek

0.15

em

0.15

Ã©g

0.14

ofire

0.13

_modules

0.13

ften

0.13

AFX

0.13

ttp

0.13

Activations Density 0.079%

pronouns and verbs that indicate user actions and instructions

No Comments

No Known Activations