INDEX

Explanations

phrases indicating the presence of placeholder pages or missing information about individuals

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

hta

-0.15

ãĤ¿ãĥ«

-0.14

itan

-0.14

_hpp

-0.14

hire

-0.14

Mahon

-0.14

abin

-0.14

Ĭ

-0.14

lain

-0.14

OMEM

-0.14

POSITIVE LOGITS

Wor

0.17

 Warner

0.16

 massaggi

0.15

urname

0.15

 ÎºÎ±Î½

0.15

 somebody

0.14

 someone

0.14

 Buchanan

0.14

.setValue

0.14

 Somebody

0.14

Activations Density 0.003%

phrases indicating the presence of placeholder pages or missing information about individuals

No Comments

No Known Activations