INDEX

Explanations

phrases related to physical impact or attacks

oai_token-act-pair · gpt-3.5-turbo

phrases indicating something has been 'hit' or impacted, particularly in a context of significant events or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rious

-0.79

iety

-0.76

naissance

-0.75

SPONSORED

-0.74

ounded

-0.73

esame

-0.73

tesy

-0.72

ilus

-0.72

asus

-0.71

acion

-0.69

POSITIVE LOGITS

 lightning

0.77

 virtue

0.75

 Hurricane

0.74

products

0.72

 mobs

0.71

 creditors

0.68

 gunfire

0.67

riot

0.65

 nature

0.64

 relegation

0.64

Activations Density 0.099%

phrases related to physical impact or attacks

phrases indicating something has been 'hit' or impacted, particularly in a context of significant events or conditions

No Comments

No Known Activations