INDEX

Explanations

computer code and programming constructs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

aklÄ±

-0.17

olie

-0.17

 rekl

-0.17

elerik

-0.16

oris

-0.16

edl

-0.15

ekl

-0.15

 Rudd

-0.15

hta

-0.15

leftright

-0.15

POSITIVE LOGITS

Jet

0.15

Dup

0.15

ones

0.15

Dup

0.14

Jub

0.14

ONES

0.14

uba

0.14

.Multi

0.14

ncy

0.14

Activations Density 0.082%

computer code and programming constructs

No Comments

No Known Activations