INDEX

Explanations

elements related to layering or stacking in a context

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

ille

-0.14

 corner

-0.14

rig

-0.14

ust

-0.14

alth

-0.14

 Reach

-0.14

-corner

-0.14

oge

-0.14

-0.13

POSITIVE LOGITS

 atop

0.42

 over

0.39

 Ð¿Ð¾Ð²ÐµÑĢÑħ

0.35

 layer

0.35

 sobre

0.33

 layers

0.31

 above

0.31

 overlay

0.30

layer

0.29

è¦Ĩ

0.29

Activations Density 0.148%

elements related to layering or stacking in a context

No Comments

No Known Activations