INDEX

Explanations

records or achievements set or broken by individuals or teams

oai_token-act-pair · gpt-3.5-turbo

phrases related to records or achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

gypt

-0.71

olitical

-0.68

xit

-0.67

sov

-0.67

arty

-0.65

markets

-0.64

farious

-0.63

enhagen

-0.62

stocks

-0.62

incinn

-0.61

POSITIVE LOGITS

 holder

1.16

 breaker

1.11

breaker

1.03

breaking

1.02

keeping

0.99

breakers

0.90

holder

0.83

 Breaker

0.82

 holders

0.82

 breaking

0.80

Activations Density 0.071%

records or achievements set or broken by individuals or teams

phrases related to records or achievements

No Comments

No Known Activations