INDEX

Explanations

names and identifiers related to specific performers and artists

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

 itself

-0.17

Å¡en

-0.15

 ìĿ´ëĬĶ

-0.14

roma

-0.14

Fizz

-0.14

ukan

-0.13

_hpp

-0.13

 partition

-0.13

 polish

-0.13

 publication

-0.13

POSITIVE LOGITS

Signature

0.18

 himself

0.18

’s

0.17

 Enterprises

0.17

bio

0.16

Bio

0.16

signature

0.16

 Himself

0.16

's

0.16

 Signature

0.16

Activations Density 0.195%

names and identifiers related to specific performers and artists

No Comments

No Known Activations