INDEX

Explanations

keywords related to positive attributes or characteristics

oai_token-act-pair · gpt-3.5-turbo

terms associated with effective communication and related professional practices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Downloadha

-0.65

bsp

-0.65

tf

-0.64

iless

-0.64

noxious

-0.62

 invaded

-0.60

Inv

-0.59

igham

-0.59

doms

-0.58

violent

-0.57

POSITIVE LOGITS

smanship

0.84

 luck

0.80

manship

0.76

 mate

0.75

 outweigh

0.75

guy

0.74

 payoff

0.74

 fodder

0.73

 handling

0.70

anke

0.69

Activations Density 0.371%

keywords related to positive attributes or characteristics

terms associated with effective communication and related professional practices

No Comments

No Known Activations