INDEX

Explanations

actions related to taking or performing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ÄĻp

-0.15

/Instruction

-0.15

ichern

-0.14

ToolTip

-0.14

antom

-0.14

 nonatomic

-0.14

agement

-0.14

IPHER

-0.14

ÑıÑĤ

-0.14

ippo

-0.14

POSITIVE LOGITS

aim

0.32

 center

0.28

 centre

0.25

 flight

0.21

Aim

0.21

 over

0.20

-to

0.20

aim

0.19

 part

0.19

 exception

0.18

Activations Density 0.046%

actions related to taking or performing

No Comments

No Known Activations