INDEX

Explanations

references to movie premieres

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "premiere"

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

nuts

-0.82

ila

-0.70

agh

-0.70

urat

-0.67

hematically

-0.67

ateg

-0.66

akia

-0.65

 Cogn

-0.65

bath

-0.64

blank

-0.64

POSITIVE LOGITS

 premiere

1.08

 screening

0.94

 airing

0.91

 episode

0.89

 premie

0.86

 Premiere

0.85

 screenings

0.85

 reel

0.83

 premier

0.80

 Episode

0.79

Activations Density 0.011%

references to movie premieres

occurrences of the word "premiere"

No Comments

No Known Activations