INDEX

Explanations

references to competitive skiing events and achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.20.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

 Cycle

-0.16

luk

-0.16

arter

-0.16

umn

-0.15

 cycle

-0.15

Rao

-0.15

arters

-0.15

 GLenum

-0.15

etz

-0.15

oyal

-0.15

POSITIVE LOGITS

Ski

0.20

ski

0.18

 poles

0.17

abay

0.16

ussy

0.16

bib

0.15

uits

0.15

TED

0.15

ailer

0.15

 skiing

0.14

Activations Density 0.007%

references to competitive skiing events and achievements

No Comments

No Known Activations