INDEX

Explanations

phrases that indicate intentions or actions aimed at facilitating or preventing specific outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

akash

-0.15

KA

-0.15

 sublic

-0.15

erts

-0.14

emos

-0.14

ÏĢÎ®

-0.14

ÙĨÚ¯

-0.14

Bab

-0.13

ulla

-0.13

POSITIVE LOGITS

avoid

0.19

 avoid

0.19

 avoiding

0.18

orum

0.17

 prevent

0.15

éģ¿

0.15

 comply

0.15

Avoid

0.15

protect

0.15

sar

0.15

Activations Density 0.194%

phrases that indicate intentions or actions aimed at facilitating or preventing specific outcomes

No Comments

No Known Activations