INDEX

Explanations

terms related to maximizing or optimizing something, such as numbers, capacities, efficiencies, or chances

oai_token-act-pair · gpt-3.5-turbo

instances of the word "maximize" and its variations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ker

-0.84

bane

-0.81

worldly

-0.76

bro

-0.71

BuyableInstoreAndOnline

-0.70

ying

-0.67

gently

-0.66

ling

-0.65

FORE

-0.64

shake

-0.64

POSITIVE LOGITS

 maximizing

1.08

 maximize

1.06

imize

1.00

imum

0.99

nesota

0.98

 efficiency

0.93

 maxim

0.90

 throughput

0.85

 flexibility

0.84

 optimize

0.84

Activations Density 0.023%

terms related to maximizing or optimizing something, such as numbers, capacities, efficiencies, or chances

instances of the word "maximize" and its variations

No Comments

No Known Activations