INDEX

Explanations

the word "whose."

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

atics

-0.15

oman

-0.14

alse

-0.14

tra

-0.14

ings

-0.14

types

-0.14

ech

-0.13

type

-0.13

ual

-0.13

POSITIVE LOGITS

oping

0.20

upon

0.19

own

0.19

/if

0.18

oped

0.18

soever

0.16

own

0.16

Own

0.16

ORA

0.15

ÑĥÐ·Ñĭ

0.15

Activations Density 0.012%

the word "whose."

No Comments

No Known Activations