INDEX

Explanations

references to shell-related terms or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

ymous

-0.17

ese

-0.16

ICC

-0.15

ethe

-0.15

edition

-0.15

ess

-0.15

ewriter

-0.15

ew

-0.15

ed

-0.15

ement

-0.15

POSITIVE LOGITS

fish

0.32

acking

0.25

acked

0.21

mates

0.20

fire

0.20

Shock

0.19

-shell

0.19

(shell

0.19

undy

0.19

ular

0.19

Activations Density 0.013%

references to shell-related terms or concepts

No Comments

No Known Activations