INDEX

Explanations

motor racing related abbreviations and terms, specifically "GP" for Grand Prix

oai_token-act-pair · gpt-3.5-turbo

references to "GP" or "GP racing"

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

ebted

-0.80

say

-0.67

Accessory

-0.64

bar

-0.64

 Jolly

-0.64

framework

-0.63

hold

-0.62

lihood

-0.61

gren

-0.61

croft

-0.60

POSITIVE LOGITS

odcast

0.87

GP

0.86

ross

0.78

ILS

0.76

Ps

0.76

ixel

0.74

OST

0.74

oline

0.74

ATES

0.73

HHHH

0.73

Activations Density 0.007%

motor racing related abbreviations and terms, specifically "GP" for Grand Prix

references to "GP" or "GP racing"

No Comments

No Known Activations