INDEX

Explanations

instances of the word "feature" or its variations, indicating a focus on presenting qualities or characteristics

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

erot

-0.16

addon

-0.15

 graz

-0.14

qv

-0.14

abyrin

-0.14

recht

-0.14

ULA

-0.14

 Thumbnail

-0.14

ian

-0.14

oper

-0.14

POSITIVE LOGITS

0.17

postcode

0.16

ŀæĢ§

0.16

erv

0.16

anmar

0.15

lah

0.15

lÃ©

0.14

opc

0.14

Lov

0.13

itÃ¤

0.13

Activations Density 0.022%

instances of the word "feature" or its variations, indicating a focus on presenting qualities or characteristics

No Comments

No Known Activations