INDEX

Explanations

actions involving holding or using items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 22-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L22R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.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

znik

-0.18

 tongues

-0.17

elige

-0.17

ometr

-0.16

hin

-0.15

erland

-0.15

tons

-0.14

 chaining

-0.14

asa

-0.14

hm

-0.14

POSITIVE LOGITS

held

0.37

 held

0.35

-held

0.32

 Held

0.30

 hold

0.24

 Holder

0.24

Hold

0.24

 holder

0.23

_holder

0.23

-holder

0.23

Activations Density 0.171%

actions involving holding or using items

No Comments

No Known Activations