INDEX

Explanations

expressions of moral or ethical judgment, particularly regarding actions deemed wrong

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 11-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L11R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.11.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

/desktop

-0.16

atic

-0.16

Ŀ

-0.15

ary

-0.15

ute

-0.14

udd

-0.14

rise

-0.14

/customer

-0.14

IELDS

-0.14

XI

-0.13

POSITIVE LOGITS

0.19

fully

0.19

zeitig

0.18

ainers

0.17

uesday

0.15

vals

0.15

IVES

0.15

/right

0.15

yntax

0.15

Activations Density 0.060%

expressions of moral or ethical judgment, particularly regarding actions deemed wrong

No Comments

No Known Activations