INDEX

Explanations

phrases indicating location and direction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 Huyá»ĩn

-0.15

Zuk

-0.14

grily

-0.14

klad

-0.14

fov

-0.13

bler

-0.13

Äįan

-0.13

.Scheme

-0.13

uw

-0.13

ungan

-0.13

POSITIVE LOGITS

 left

0.91

 right

0.90

left

0.75

 Left

0.73

 Right

0.71

right

0.71

å·¦

0.70

 LEFT

0.69

 RIGHT

0.69

Left

0.68

Activations Density 0.201%

phrases indicating location and direction

No Comments

No Known Activations