INDEX

Explanations

words indicating provocation or prompting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

Dit

-0.15

andro

-0.15

mony

-0.15

ê³

-0.14

spl

-0.14

ificio

-0.14

ÑĥÑĤÐ¸

-0.14

erp

-0.14

squ

-0.14

ÑĢÑĥÐ´

-0.14

POSITIVE LOGITS

ost

0.30

iso

0.25

oking

0.25

ocation

0.23

isos

0.22

ably

0.21

insi

0.20

okes

0.19

isory

0.18

okers

0.17

Activations Density 0.004%

words indicating provocation or prompting

No Comments

No Known Activations