INDEX

Explanations

complex relationships and properties within abstract concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 7-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L7R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.7.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

olicited

-0.16

 Flake

-0.15

yas

-0.15

yer

-0.14

ãĥ¼ãĥª

-0.14

Boot

-0.14

ernaut

-0.14

aret

-0.14

OLLOW

-0.14

ne

-0.14

POSITIVE LOGITS

zzo

0.17

 ÙĩÙħÛĮÙĨ

0.17

eo

0.16

IFA

0.15

imb

0.14

__$

0.14

BCM

0.14

 Ð³Ð¾ÑģÐ¿

0.14

ieÅ¼

0.13

deo

0.13

Activations Density 0.088%

complex relationships and properties within abstract concepts

No Comments

No Known Activations