INDEX

Explanations

markers and elements related to document formatting or structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

Ð³Ð¾Ð´

-0.15

quette

-0.15

ulet

-0.15

 Horny

-0.14

Î´Îµ

-0.14

imid

-0.14

æĿī

-0.14

 Spicer

-0.14

 Monter

-0.14

Ã¨m

-0.14

POSITIVE LOGITS

odia

0.14

trl

0.14

 hull

0.14

Locale

0.14

DonaldTrump

0.14

abbr

0.13

uth

0.13

 micro

0.13

cerr

0.13

Activations Density 0.155%

markers and elements related to document formatting or structure

No Comments

No Known Activations