INDEX

Explanations

terms associated with the concept of role models

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

ched

-0.17

Mov

-0.17

ernals

-0.16

ropolis

-0.15

 Policies

-0.15

Å¡etÅĻ

-0.15

nis

-0.15

ramento

-0.15

iano

-0.15

ÑĢÐ°Ð¼

-0.15

POSITIVE LOGITS

playing

0.21

ystone

0.20

(Role

0.18

 revers

0.17

(role

0.17

 reversal

0.16

 ROLE

0.16

xes

0.16

-playing

0.15

stown

0.15

Activations Density 0.009%

terms associated with the concept of role models

No Comments

No Known Activations