INDEX

Explanations

references to "the" along with other related words indicating context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

regon

-0.18

agen

-0.15

eldon

-0.15

urg

-0.15

Qed

-0.15

essim

-0.14

 downstream

-0.14

 Whites

-0.14

åº

-0.14

Fir

-0.14

POSITIVE LOGITS

ÙĪÙĨØ¯

0.16

orton

0.15

ApiClient

0.15

è¸

0.15

fcc

0.14

 GmbH

0.14

dj

0.14

YRO

0.14

?family

0.14

.camel

0.13

Activations Density 0.092%

references to "the" along with other related words indicating context

No Comments

No Known Activations