INDEX

Explanations

entity names and terms associated with culture

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

activex

-0.16

'gc

-0.16

ucch

-0.15

agine

-0.15

ablish

-0.15

imum

-0.14

ajor

-0.14

azor

-0.14

hack

-0.14

rego

-0.14

POSITIVE LOGITS

(B

0.17

 Bound

0.16

sev

0.15

avra

0.15

/B

0.15

=B

0.15

shelf

0.15

 trivia

0.15

bul

0.14

 ball

0.14

Activations Density 0.230%

entity names and terms associated with culture

No Comments

No Known Activations