INDEX

Explanations

references to educational institutions, particularly colleges

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

ties

-0.16

reas

-0.16

ness

-0.16

-serif

-0.16

ÑĬ

-0.15

uzey

-0.15

rial

-0.15

nis

-0.15

dem

-0.15

asil

-0.15

POSITIVE LOGITS

/un

0.19

yard

0.18

ofs

0.17

-level

0.16

sein

0.16

-wide

0.15

-educated

0.15

wide

0.15

pond

0.14

boy

0.14

Activations Density 0.025%

references to educational institutions, particularly colleges

No Comments

No Known Activations