INDEX

Explanations

instances of withholding or emotional withholding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

lou

-0.17

echa

-0.15

ERVED

-0.15

 thirst

-0.14

 Grad

-0.14

jar

-0.14

 Wilde

-0.14

Ð´ÑĢÐµÑģ

-0.14

ita

-0.13

 round

-0.13

POSITIVE LOGITS

ocre

0.20

)((((

0.17

_iff

0.15

angered

0.15

ImageContext

0.15

opy

0.15

posables

0.15

Äįel

0.15

warts

0.15

âĢ¦↵↵↵

0.15

Activations Density 0.014%

instances of withholding or emotional withholding

No Comments

No Known Activations