INDEX

Explanations

the word "whole."

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "whole" as it relates to completeness or entirety

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

osponsors

-0.73

KER

-0.71

akers

-0.69

ANS

-0.68

licts

-0.64

friends

-0.64

ics

-0.63

rets

-0.63

zynski

-0.63

cl

-0.62

POSITIVE LOGITS

 thing

1.02

heartedly

0.99

hearted

0.96

 darn

0.87

 globe

0.84

gam

0.84

 ordeal

0.82

 spectrum

0.81

 damn

0.80

 world

0.79

Activations Density 0.032%

the word "whole."

references to the concept of "whole" as it relates to completeness or entirety

No Comments

No Known Activations