INDEX

Explanations

terms related to weakness or vulnerability

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

fully

-0.18

iest

-0.16

pill

-0.15

ertz

-0.15

Ð°ÑĢÐ°Ð¼

-0.15

ÑĥÑĤÑĤÑı

-0.15

rijk

-0.14

ĳ¸

-0.14

Äįer

-0.14

ech

-0.14

POSITIVE LOGITS

 spot

0.26

 spots

0.25

ling

0.23

ly

0.22

lings

0.22

spot

0.22

 weak

0.20

_spot

0.20

å¼±

0.20

 Spot

0.20

Activations Density 0.023%

terms related to weakness or vulnerability

No Comments

No Known Activations