INDEX

Explanations

complex identifiers or variables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

Fuck

-0.16

fuck

-0.16

 FUCK

-0.15

 fucked

-0.15

 fuck

-0.15

 Wrestling

-0.15

 fucking

-0.14

 Gundam

-0.14

 Fucking

-0.14

 fucks

-0.14

POSITIVE LOGITS

 Hidden

0.25

HO

0.23

 hidden

0.22

 Collector

0.21

 puzzles

0.21

Hidden

0.20

 games

0.20

 gameplay

0.19

CE

0.19

 scenes

0.19

Activations Density 0.007%

complex identifiers or variables

No Comments

No Known Activations