INDEX

Explanations

terms related to mathematical definitions and theorems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ÙĬÙĥØ§

-0.16

eya

-0.15

 Elli

-0.15

Ø¨Ø§Ø±

-0.15

oden

-0.15

 denotes

-0.14

 denote

-0.14

den

-0.14

ëĭ

-0.14

 accents

-0.14

POSITIVE LOGITS

 saying

0.25

 valid

0.24

 stating

0.24

 holds

0.23

 stated

0.23

 established

0.23

 asserting

0.22

 holding

0.21

 states

0.21

 Holds

0.21

Activations Density 0.150%

terms related to mathematical definitions and theorems

No Comments

No Known Activations