INDEX

Explanations

expressions emphasizing the term "so" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pane

-0.16

nty

-0.15

eous

-0.15

more

-0.15

apolis

-0.15

Ð¼Ð¾ÑĢ

-0.15

phan

-0.14

bies

-0.14

gram

-0.14

incy

-0.14

POSITIVE LOGITS

-called

0.42

 much

0.36

oo

0.33

apy

0.33

oooo

0.32

iled

0.32

ooo

0.30

oth

0.29

ars

0.28

aks

0.28

Activations Density 0.055%

expressions emphasizing the term "so" in various contexts

No Comments

No Known Activations