INDEX

Explanations

variations of the token "stre" and its related forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 Middleton

-0.15

’ÑıÐ·

-0.15

arrant

-0.14

ammers

-0.14

ãĤĵãģ©

-0.14

rium

-0.14

 Midnight

-0.14

 midnight

-0.14

Ã¤Ã¤n

-0.14

 Likes

-0.14

POSITIVE LOGITS

ngthen

0.27

 stre

0.26

 Stre

0.20

NGTH

0.20

 stretch

0.19

ength

0.19

/stretch

0.18

elight

0.18

etwork

0.18

ngth

0.18

Activations Density 0.004%

variations of the token "stre" and its related forms

No Comments

No Known Activations