INDEX

Explanations

references to innocence or themes of being innocent in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 7-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L7R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.7.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

olon

-0.18

ICO

-0.16

asu

-0.15

/xhtml

-0.15

aN

-0.15

recision

-0.15

ongan

-0.15

istrovstvÃŃ

-0.15

abo

-0.15

ÙĨÚ¯

-0.14

POSITIVE LOGITS

 innoc

0.20

 Innoc

0.19

 innocent

0.19

ubs

0.18

 innocence

0.18

ves

0.16

endale

0.15

ules

0.15

arya

0.15

 bystand

0.14

Activations Density 0.018%

references to innocence or themes of being innocent in various contexts

No Comments

No Known Activations