INDEX

Explanations

phrases related to support and rescue in various contexts

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

à¤¿à¤£

-0.18

-0.17

ETO

-0.15

 Interr

-0.15

heimer

-0.15

éļ

-0.14

 blindness

-0.14

æ¿Ģ

-0.14

 gÃ³

-0.14

etimes

-0.14

POSITIVE LOGITS

fal

0.38

 failing

0.36

sag

0.35

 dying

0.33

 flag

0.30

ailing

0.30

 struggling

0.28

dec

0.27

lag

0.26

 fading

0.25

Activations Density 0.409%

phrases related to support and rescue in various contexts

No Comments

No Known Activations