INDEX

Explanations

elements related to fragments or sections of texts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

 towns

-0.16

ulp

-0.15

 town

-0.14

ToLocal

-0.14

IndexOf

-0.14

unge

-0.13

weed

-0.13

 Hindered

-0.13

hci

-0.13

ache

-0.13

POSITIVE LOGITS

 portion

0.20

areas

0.17

area

0.17

 area

0.17

 portions

0.16

-area

0.15

inium

0.15

 section

0.15

 Powers

0.15

 areas

0.15

Activations Density 0.246%

elements related to fragments or sections of texts

No Comments

No Known Activations