INDEX

Explanations

expressions related to anticipation or expectation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

 look

-0.23

 looking

-0.22

 Look

-0.21

 Looking

-0.21

looking

-0.21

look

-0.19

 looks

-0.19

Looking

-0.19

Look

-0.18

 LOOK

-0.18

POSITIVE LOGITS

 forward

0.30

forward

0.24

 FORWARD

0.23

-forward

0.21

 forwarding

0.21

 Forward

0.20

Forward

0.20

 forwarded

0.19

fwd

0.19

.forward

0.18

Activations Density 0.015%

expressions related to anticipation or expectation

No Comments

No Known Activations