INDEX

Explanations

adverbs, particularly those ending in 'ly' and various forms of 'mente'

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L1R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.1.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

ovky

-0.15

RIPT

-0.15

gens

-0.15

oin

-0.14

aged

-0.14

oved

-0.13

loit

-0.13

 chatte

-0.13

ÏĦÎµÏģ

-0.13

Ä±rÄ±

-0.13

POSITIVE LOGITS

 cond

0.15

esa

0.15

cal

0.14

Flash

0.14

ocket

0.14

jez

0.14

UF

0.14

ucer

0.14

IME

0.14

heim

0.13

Activations Density 0.083%

adverbs, particularly those ending in 'ly' and various forms of 'mente'

No Comments

No Known Activations