INDEX

Explanations

phrases related to explosive or impactful actions

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

ummings

-0.18

ully

-0.16

ussia

-0.15

Å¡ov

-0.15

oret

-0.15

ucer

-0.15

Ä±lÄ±Ã§

-0.15

uzzle

-0.14

uite

-0.14

 oversh

-0.14

POSITIVE LOGITS

torch

0.27

 kisses

0.23

 blew

0.22

 away

0.22

 Fuse

0.21

 apart

0.21

Fuse

0.20

 fuse

0.20

away

0.20

 minds

0.19

Activations Density 0.010%

phrases related to explosive or impactful actions

No Comments

No Known Activations