INDEX

Explanations

references to publication details and bibliographic information

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

 favored

-0.17

rl

-0.17

 favor

-0.15

 honor

-0.15

 Buff

-0.15

Ù¾Ø³

-0.15

_buff

-0.15

 honors

-0.15

 buffalo

-0.15

OWL

-0.15

POSITIVE LOGITS

 Rout

0.19

Hod

0.18

 Guardian

0.18

‘

0.18

 Penguin

0.17

Britain

0.17

 Thames

0.17

enguin

0.16

BBC

0.16

Activations Density 0.143%

references to publication details and bibliographic information

No Comments

No Known Activations