INDEX

Explanations

events or ceremonies named "Awards."

oai_token-act-pair · gpt-3.5-turbo

references to various award ceremonies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Anon

-0.77

lessly

-0.71

ths

-0.70

lessness

-0.70

gun

-0.69

JM

-0.68

Beg

-0.67

ohyd

-0.66

ulia

-0.66

ufact

-0.66

POSITIVE LOGITS

 Awards

1.10

 nominations

0.99

 Oscars

0.94

 ceremony

0.88

 awards

0.83

 nominees

0.83

creen

0.81

 ceremonies

0.79

 prizes

0.77

hips

0.76

Activations Density 0.017%

events or ceremonies named "Awards."

references to various award ceremonies

No Comments

No Known Activations