INDEX

Explanations

words and phrases related to emotional reactions and communications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.20.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

 Panic

-0.16

 panic

-0.16

panic

-0.15

åĭ¤

-0.15

.questions

-0.15

iment

-0.15

 Asking

-0.14

ilia

-0.14

ember

-0.14

 panicked

-0.14

POSITIVE LOGITS

 hurt

0.18

 Dipl

0.17

venes

0.15

 receipts

0.15

ÚĺÛĮ

0.15

 confrontation

0.15

angered

0.15

apyrus

0.15

okino

0.14

 toxic

0.14

Activations Density 0.171%

words and phrases related to emotional reactions and communications

No Comments

No Known Activations