INDEX

Explanations

the word "only" in various contexts indicating limitations or exclusivity

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

 already

-0.18

 Already

-0.16

already

-0.16

Already

-0.15

 neither

-0.15

iets

-0.15

ajo

-0.15

 ÑĥÐ¶Ðµ

-0.14

alth

-0.14

 kein

-0.14

POSITIVE LOGITS

 ever

0.33

 margin

0.26

-ever

0.25

 partially

0.24

 really

0.23

ever

0.23

 recently

0.23

 EVER

0.21

 truly

0.20

 Partial

0.19

Activations Density 0.057%

the word "only" in various contexts indicating limitations or exclusivity

No Comments

No Known Activations

the word "only" in various contexts indicating limitations or exclusivity

No Comments

No Known Activations