INDEX

Explanations

phrases related to conditions and subject to approval

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

uo

-0.15

aty

-0.15

ÃŁ

-0.15

zug

-0.15

å¼ĺ

-0.14

arp

-0.14

 deserted

-0.13

 INCIDENTAL

-0.13

charts

-0.13

lessness

-0.13

POSITIVE LOGITS

 conditions

0.24

 approval

0.23

 availability

0.22

conditions

0.21

 Conditions

0.20

 certain

0.19

 receipt

0.19

 approvals

0.19

Conditions

0.18

 final

0.18

Activations Density 0.054%

phrases related to conditions and subject to approval

No Comments

No Known Activations